This paper extends research linking shopping behaviour to environmental factors through changes in emotional states. With time fixed or variable during a simulated shopping experiment, shoppers were exposed to music varying by degree of familiarity. Afterward, subjects reported their perceptions of shopping duration, their emotional states, and their merchandise evaluations. Analyses revealed that individuals reported themselves as shopping longer when exposed to familiar music but actually shopped longer when exposed to unfamiliar music. Shorter times in the familiar music condition were related to time misperceptions. Although emotional states affected product evaluations, these effects were not easily related to the music manipulations.
Phil Kotler introduced the view that retail environments create atmospheres that affect shopping behavior in the Journal of Retailing in 1973. Although a special issue devoted to the subject (Journal of Retailing, Winter 1974) followed shortly, the area did not otherwise receive much attention. Donovan and Rossiter (1982) revived interest by suggesting that environmental psychology, especially Mehrabian and Russell’s (1974) PAD framework could be used productively to research store environments. Researchers subsequently analysed retail shopping behaviour with this framework and found significant relationships between emotional states and factors such as time spent in the store, propensity to make a purchase, and satisfaction with the experience (e.g., Sherman and Smith 1987; Dawson, Bloch and Ridgway 1990; Kellaris and Kent 1993; Yalch and Spangenberg 1993).
For example, Sherman and Smith (1987) interviewed shoppers immediately after a purchase and solicited responses regarding their shopping experience, their mood, and demographic characteristics. They found positive relationships between shoppers’ reported mood and how favourably shoppers perceived the store, how many items they purchased, and how much time they spent in the store. Unfortunately, as a correlational study, cause and effect was indeterminate. Given the extensive psychological research showing that individuals partially judge their emotional states by their behaviour (e.g., Bem 1972; Schachter and Singer 1962), the causal role of emotional reactions to environmental factors in determining shopping behaviours remains uncertain.
Experimental control of environmental factors provides a better, more controlled, test of atmospheric factors and the moderating role of subsequent emotional states on shopping behaviour. Several studies have manipulated atmospheric factors such as crowding (e.g., Eroglu and Harrell 1986; Eroglu & Machleit 1990; Harrell, Hutt, and Anderson 1980; Hui and Bateson 1991), colors (e.g., Bellizzi, Crowley and Hasty 1983), music (e.g., Kellaris and Altsech 1992; Milliman 1982, 1986; Yalch and Spangenberg 1988; 1993), and olfactory cues (Spangenberg, Crowley, and Henderson 1996) and tested their effects on shopping behaviors like satisfaction with the shopping experience (e.g., Eroglu & Machleit 1990; Bellizzi et al. 1983), purchase quantity (Milliman 1982; 1986), shopping times (Kellaris & Altsech 1992, Milliman 1982; 1986, Yalch and Spangenberg 1988), and intention to visit the store again (Spangenberg et al. 1996). However, with few exceptions (i.e., Hui and Bateson 1991; Yalch and Spangenberg 1988), these studies have not examined the emotional states of pleasure, arousal, and dominance postulated by Donovan and Rossiter (1982) as factors mediating the effect of retail environments on behaviour.
The current article reports a study of retail shopping that includes all three aspects considered important in studies of environmental psychology: environmental stimuli, emotional reactions, and shopping behaviour. Shoppers listened to store music varying in its familiarity while they examined various articles of outdoor clothing. Our objective was to learn whether store music influenced shoppers’ emotional states and, if so, whether these emotional states subsequently affected shopping behaviour. Specific influences measured included time spent shopping, perception of the amount of shopping time, and actual product evaluations.
ENVIRONMENTAL PSYCHOLOGY IN RETAILING
Mehrabian and Russell (1974) developed a framework for analysing the effects of environments on individuals, emphasizing the role of nonverbal responses to environmental factors as a major determinant of behaviour. Related to Bitner’s (1992) exploration of how physical environments might affect both employees and customers and Donovan and Rossiter’s (1982) PAD framework, our Figure illustrates how store environments might influence shopping behaviour through mediating emotional states. The store environment contains various stimuli that might be perceived by the customer’s senses; each stimulus offers many options with regard to variability. For example, store music varies by volume, tempo, pitch and texture and by the specific songs played (see Bruner 1990 for a lengthy discussion of music as a marketing stimulus). In addition, factors can be combined to create unique atmospheres. To project an upscale image, a manager might choose classical music, subdued colours, elegant perfumes, cool temperatures, sparsely displayed merchandise, and low lighting.
Mehrabian and Russell’s (1974) framework specifies that individuals react to their environment along at least three dimensions: Pleasure, Arousal, and Dominance (PAD). The first is an effectual reaction, labelled Pleasure-Displeasure. This entails whether individuals perceive the environment as enjoyable or not enjoyable. For example, playing popular songs should enhance shoppers’ enjoyment, whereas unpopular music might diminish it. The second dimension relates to Arousal. It assesses how much the environment stimulates the individual. Milliman (1982) found that playing slow music resulted in slower customer movement through a supermarket relative to no music or fast music; this result is perhaps attributed to a decrease in arousal. Relatedly, Kellaris and Kent (1993) found main effects for music tempo on subject arousal in a laboratory study. The third dimension is Dominance which concerns whether individuals feel dominant (in control) or submissive (under control) in the environment. An early study by Babbitt (1878) reported that violent patients became more violent when placed in a red colour room but less violent when in a blue room. This is consistent with the finding that individuals associate the colour red with active, assertive, and rebellious moods (Aaronson 1970), whereas they associate blue with sedate tranquillity and a suppression of feelings (Schaie and Heiss 1964).
The PAD model has been tested in both retailing and non-retailing environments. Sherman and Smith (1987) combined Pleasure and Arousal in a retailing study as a unidimensional mood scale but this practice is inconsistent with the multi-dimensional theory underlying the measures (Russell 1980). Donovan and Rossiter (1982) treated the three dimensions as distinct factors and reported that shopping behaviours were related to measures of Pleasure and Arousal but not Dominance. In non retailing environments, research generally supports the PAD model with the exception that Dominance is sometimes not a major factor or simply not measured (Russell 1980). It remains unclear whether difficulties in identifying behaviours associated with Dominance reflect its small influence on behaviour or the need for improved measurement.
The last element of the framework presented in the Figure is a taxonomy of possible behavioural reactions to the environment (cf. Donovan & Rossiter 1982). Although a variety of shopping behaviors can be affected in many ways by environmental factors, Mehrabian and Russell (1974) suggested use of an approach-avoidance paradigm. Thus, environments could be constructed to encourage or discourage approach behaviours. For example, bright colours might encourage individuals to enter a fast food restaurant, whereas uncomfortable seating might discourage long stays. These approach-avoidance behaviours can be grouped into four categories based on the type of behaviour: time, exploration, communication, and satisfaction. Table 1 summarizes these four types of behaviours applied to retail shopping. The current research focuses on how much time consumers choose to spend shopping as a function of the type of music played in the store.
Taxonomy of Retailing Shopping Behaviors
1. A desire to physically stay in or to get out of the environment. This relates to the decision to shop or not to shop at the store. It also might relate to the length of time spent in the store. Presumably, attractive in-store environments build store traffic and encourage individuals to linger in the stores. However, it might be difficult to accomplish both simultaneously. For example, bright yellow might attract customers but prolonged exposure might be over stimulating resulting in a quick departure. [TIME]
2. A desire or willingness to explore the environment. This relates to how much area of the store is covered. Stores are sometimes designed like caves with areas not easily seen from a single location to encourage shoppers to walk around. Having hidden sale displays might further reward and encourage shoppers to explore the entire store. Dark colors might suggest mystery and entice exploration.[EXPLORATION]
3. A desire or willingness to communicate with others in the environment. This would be particularly important in retail stores in which customers must rely on the sales staff to describe and explain the items in the store. Neutral colors might minimize status differences and thereby encourage greater social interaction. [COMMUNICATION]
4. The degree of enhancement or hindrance of performance and satisfaction with task performance. This relates to factors such as the ability of customers to locate what they want, to purchase the item with a minimal wait in line, and to transport easily the item from the store to their car. Express lines and carryout help facilitate this aspect of grocery shopping. [SATISFACTION]
MUSIC AND TIME PERCEPTIONS
Time is an important factor in retail shopping, partially because studies show a simple correlation between time spent shopping and shopping and amount purchased (cf. Milliman 1982). Also, time is argued to be as much a constraint on consumption as money and that predictions that individuals would have more time at their discretion in the future than in the past have not proven to be true (Berry 1979). For example, dual career families with children coping with the transportation difficulties in dense metropolitan areas may feel intense time pressure when shopping. Consequently, it is reasonable to expect individuals to budget their time, including shopping times, and to be concerned when they believe they are spending too much time in a store. People simply don’t enjoy waiting too long or wasting time. Hornik (1984), for example, reported that shoppers overestimated their waiting time less when they reported a high level of shopping enjoyment relative to other activities. Thus, retailers would be prudent to minimize perceived as well as actual time spent shopping for their patrons.
Field research by Yalch and Spangenberg (1988) suggested that music affects shopping times. In their study, clothing store shoppers were exposed either to a youth-oriented foreground music or adult-oriented background music. Interviews with shoppers as they were exiting the store revealed that younger shoppers felt they had shopped longer when exposed to background music, whereas older shoppers felt they had shopped longer when exposed to foreground music. Unfortunately, actual shopping times were not observed so it could not be determined if individuals shopped longer, merely thought that they did, or a combination of both factors.
Milliman’s (1982) results suggest that music affects actual shopping times. In his study, restaurant patrons were exposed to either fast or slow tempo music. Individuals tended to stay longer when listening to the slow music compared to the fast music. The additional time did not result in any greater expenditure on food but did lead to an increase in the amount spent on drinks. No surveys were administered so it is not known whether the restaurant patrons were aware that they were spending more time or not.
Kellaris and Altsech (1992) supported the belief that music affects time perceptions. Individuals listened to original music composed in a light popular style, lasting 180 seconds. Its loudness varied from being either loud or soft, corresponding to levels associated with foreground or background music. Afterward, male and female subjects estimated song length. The results indicated no differences in perceived duration for males but that females perceived the loud music as lasting much longer than the soft music. These results, however, are not directly applicable to retailing because individuals were listening to music and not shopping. Further, exposure time was controlled for all individuals so it is not known if the volume of music would have affected self-determined listening times.
Although not conducted in a retail setting, perhaps the most surprising recent finding regarding the effect of music on time perception was that of Kellaris and Cox (1992). In a laboratory study, they found that modality affected listeners estimates of time period duration in a manner contradictory to the conventional wisdom of “time flies when your having fun.” Perceived duration of time was longest for subjects exposed to positively valance music and shortest for negatively valence music. Thus, contrary to popular belief, in this experiment, time did not fly when the time interval was filled with an affectively positive musical selection.
In summary, research suggests a relationship between characteristics of environmental music and both the actual and perceived amount of time devoted to a task; however, no study has looked at both effects simultaneously. In addition, most research on store music has varied actual qualities of the music (e.g., tempo), and not consumer perceptions of it (e.g., familiarity or liking). Both approaches are useful because retailers may select music based on its listener responses and familiarity as well as its other qualities. Additionally, listener responses are frequently studied in advertising contexts (e.g., Gorn 1982; Kellaris, Cox and Cox 1992), offering the potential for a convergence in theory and results.
An experiment was conducted to learn how individuals engaged in a shopping activity might be affected by environmental music varying in its perceived familiarity to the shoppers. Fontaine and Schwalm (1979) reported that, relative to no music or unfamiliar music, playing familiar music increased subjects’ arousal level, vigilance in detecting visual signals, and mitigated an expected decline in vigilance as the length of the session increased. Thus, it is expected that individuals listening to familiar music will be more aroused and spend more time shopping than individuals listening to unfamiliar music. This was tested in the current study by observing the length of time individuals shopped when given a choice as to how much time to spend shopping.
A second issue concerns the relationship between music and perceived time. Perceptions of greater duration appear to be associated with awareness of the environment and activities occurring in it (Zakay, Nitzan & Glicksohn 1983). Because listening to familiar music seems to cause individuals to be more vigilant compared to listening to unfamiliar music (Fontaine and Schwalm 1979), shoppers were expected to perceive more time passing. Thus, it was expected that for shoppers who had a fixed amount of time to shop, individuals would estimate that more time transpired when listening to familiar music compared to unfamiliar music.
A third issue concerns the effect of music on product evaluations. Because individuals were expected to feel more comfortable in an environment featuring familiar atmospheric elements than unfamiliar ones, their product evaluations were expected to be more favorable when familiar music was played compared to unfamiliar music.
The final research issues pertain to the mediating role of the emotional measures of pleasure, arousal, and dominance. Individuals were expected to evidence more arousal, more pleasure, and a greater sense of dominance when listening to familiar music than unfamiliar music. These emotional responses, in turn, were expected to account for most of the effects of music on perceived and actual shopping times and product evaluations.
A 2×2 factorial experiment was conducted to determine how time spent shopping might be affected by the type of music being played in the environment. One factor varied was the music being played while respondents shopped. Half heard familiar “contemporary” music and the other half unfamiliar “easy listening” music. The other factor was control over the time spent shopping. Half the subjects were given a fixed amount of time to shop whereas the other half could shop as long as they wanted. Shopping time was controlled in order to separate perceptual effects from actual effects. This necessitated using a simulated shopping experience instead of naturally occurring shopping. Dependent measures included the amount of time actually spent shopping, the amount of time perceived to have been spent, and product evaluations. In addition, respondents completed a modified version of Donovan and Rossiter’s (1982) measures of emotional responses to environmental stimuli.
Seventy-one individuals were recruited from marketing classes to participate in a new product evaluation task. In groups of three to six, subjects entered a classroom set up to appear like a clothing store. Ten articles of outdoor outerwear and equipment were displayed on tables and one of the two types of music was provided by a concealed tape recorder. Each subject completed a questionnaire while examining three items they chose from those on display. Half of the subjects were given a fixed amount of time to complete the task. The other half had an unlimited amount of time. Debriefing occurred after all groups had participated.
Music Familiarity. Two tapes were provided by a national supplier of environmental music. One tape consisted of familiar music, mostly top 40’s song designed to appeal to college-aged individuals. The other tape was unfamiliar (to our subjects) music, older songs played in an instrumental form. Although songs on the unfamiliar music tape were less well-known to subjects, it was still likely that they would find it enjoyable as this music was designed by the providing firm to appeal to a broad cross-section of the population. Thus, it was expected that subjects would express little or no difference in liking for the two types of music; only differences in familiarity should be found.
Time Control. The other manipulation in the experiment involved control over the amount of time available to complete the product evaluation task. Half of the subjects were given the opportunity to take as much time as they wanted. The other half were given exactly eleven minutes for the task; pretests determined that this was a reasonable amount of time to examine the merchandise. The purpose of this manipulation was to distinguish actual and perceptual time effects. If subjects spent more time when listening to familiar music, this would be evident in the variable time condition. Efforts to isolate a perceptual effect without controlling time would be hindered by the possibility that perceptual differences are related to duration. For example, individuals may misestimate time more as its duration increases. With this study’s design, music’s separate effect on perceptions could be observed in the fixed time condition.
The questionnaire was included in a shopping booklet. The first page was a cover sheet providing a description of the task as an evaluation of several proposed new types of outdoor equipment and clothing from a local manufacturer. Next, subjects answered a series of background questions related to subjects’ experience and interest in outdoor activities. This was largely used to disguise the study as concerning outdoor products. The next three pages were identical and provided evaluation questions for three of the items on display. Subjects were instructed to assume that they were shopping for outdoor equipment and to select items of interest. The subjects wrote the name of the item at the top, evaluated it on seven bipolar adjective scales consisting of items like Lowest performance/Highest performance and Not at all stylish/Very stylish. In addition, subjects indicated their likelihood of buying the item, what they thought was an appropriate price, and the maximum price they would pay. Finally, they were provided with several blank lines and encouraged to write down any other thoughts that they might have. This was intended to encourage additional scrutiny of the items and provide an opportunity for shoppers to expand or shorten the task.
Subjects informed the experimenter when they had completed their shopping experience. At this point, they estimated the amount of time they had spent doing their product evaluations by marking an “X” on a dashed line with 60 dashes and points marked at 0, 5, 10, 15 and 20 minutes. The next to last page of the questionnaire consisted of a series of bipolar adjectives assessing the emotional responses (PAD) of the subjects to the simulated retail environment. The last page had several scales to determine the subjects’ reactions to the room (sense of being crowded or not, temperature, lighting) and the music (liked/disliked and usually listened to/rarely listened to). These scales were used to test the construct validity of the music manipulation.
Music Manipulation. The subjects’ evaluations of the familiar and unfamiliar music at the end of the questionnaire were used to test whether their perceptions corresponded to the experimenters’ expectations. T-tests revealed no difference in liking of the music (familiar = 3.9, unfamiliar = 4.0, t(69) < 1.0), but a significant difference in familiarity (familiar = 3.9, unfamiliar = 1.9, t(69) = 4.5, p < .001). Thus, our expectation that liking would not differ between the two types of music was supported and the familiarity manipulation was successful.
Other Environmental Factors. Evaluations of the retail setting in terms of lighting and temperature revealed no differences between the music conditions (all t’s < 1). Subjects perceived the room as somewhat more comfortable when listening to familiar music than unfamiliar music (means = 3.4 vs. 2.7, t(69) = 1.6, p < .15).
Effects of Music on Time. The effects of music on actual and perceived shopping times were assessed in several ways (see Table 2 for cell means). First, variation in actual time was examined by comparing subjects in the variable time conditions who were exposed to familiar and unfamiliar music. More time was spent when the unfamiliar music was being played, (t(37) = 1.7, p < .1). This is contrary to our expectation that familiar music would encourage longer shopping.
To test the effects on perceived time, a comparison was made between unfamiliar and familiar music groups when shopping time was restricted to eleven minutes. The results were as expected (Table 2). Subjects perceived themselves as spending longer when exposed to the more familiar music compared to the less familiar music (t(28) = 2.02, p < .05).
Finally, the effect of music on perceived shopping time was analysed in the variable time condition (Table 2). There was no difference in the perceptions of how much time was spent shopping between the two music conditions (t(37) < 1).
Effects of Music on Nonverbal Responses. Responses to the seventeen bipolar adjective scales adaptive from Donovan and Russell’s (1982) scales were factor analysed, resulting in three clearly identifiable factors. The Pleasure dimension consisted of contented-depressed, happy-unhappy, satisfied-unsatisfied, annoyed-pleased, bored-interested, hopeful-despairing. Responses to these items were summed and considered a measure of positive and negative mood during the task (Chronbach’s = .88). The second factor was labelled Arousal and was composed of the items relaxed-stimulated, calm-excited, aroused-unaroused, and frenzied-sluggish, (Chronbach’s = .70). The third dimension was Dominance and included responses to three items, controlled-in control, dominant-submissive, and influenced-influential (Chronbach’s = .52).
Differences in emotional responses due to the type of music and variations in the amount of time allowed for shopping are reported in Table 2. Analysis of variance indicated a significant interaction of time control and music familiarity on the Pleasure dimension (F[1,65] = 5.4, p < .03). Analyses were run separately for the fixed and variable time groups, revealing a significant effect of music only in the fixed time condition. Contrary to expectations, individuals reported greater pleasure when listening to unfamiliar music compared to familiar music (t(28) = 2.2, p < .05).
There was also a significant interaction effect between music familiarity and control of time (F[1,65] = 4.2, p < .05) for the Arousal measure. Separate analyses for the two time conditions revealed no effect of music in the fixed time condition. In the variable time condition, individuals reported greater arousal when listening to the unfamiliar music compared to familiar music (t(37) = 1.9, p < .1).
For the Dominance measures, although there was a significant overall effect of music (F(1,64) = 4.8, p < .05), separate analysis by control of time revealed that the effect was significant only in the fixed time condition. Here, individuals reported greater sense of dominance when listening to the unfamiliar music compared to the familiar music (t(29) = 3.0, p < .01).
Mediating Role of Emotions on Shopping Times. The mediating effect of the nonverbal responses on actual and perceived shopping times was tested using analysis of covariance. The statistical procedure used the three emotional states as covariates and type of music as the treatment factor. For subjects in the variable time condition, the results were that actual time was related to Arousal (F(1,33) = 6.6, p < .05) and that once this effect had been considered the effect of music was no longer significant (p > .3). For subjects in the controlled time condition, perceived shopping time was not significantly related to any of the covariates.
Mediating Role of Emotions on Product Evaluations. Although subjects could evaluate three products, some subjects in the fixed time condition did not have enough time to complete their evaluation of the third product and some in the variable time condition chose not to evaluate a third product. Therefore, only responses to the seven bipolar adjective scales for the first two products were summed and averaged to develop a measure of overall product evaluations. The effects of type of music and control of time on these evaluations was tested using analysis of covariance with Pleasure, Arousal, and Dominance as the covariates for both the fixed and variable time conditions. Because there was a significant interaction between type of music and control of time (F(1,53) = 4.6, p < .05), separate analyses were conducted for the fixed and variable time conditions. For the fixed time condition, the results were that product evaluations were affected by Pleasure, (F(1,53) = 5.5, p < .05) and music (F(1,18) = 5.7, p < .05). Shoppers evaluated the two products more favourably when listening to familiar music compared to unfamiliar music. In the variable time condition, only Arousal (F(1,32) = 5.2, p < .05) was significantly related to product evaluations. Correlational analysis across all groups revealed that product evaluations were positively related to Pleasure (r = .33) but negatively related to Arousal (r = -.25). Music had no significant effect in the variable time condition.
The results of this study support the belief that shopping time is affected by a retail environmental factor like store music. Individuals who had a choice as to the duration of their shopping experience shopped longer when listening to less familiar music compared to more familiar music. This difference appeared attributable to differences in emotional responses to the two types of music. Individuals reported being more aroused while listening to the unfamiliar music compared to the familiar music. Once the effect of arousal on shopping times was considered, other reactions to music familiarity (either measured or unmeasured) did not have an effect on actual shopping times.
These results are counter to the expectation that listening to familiar music would encourage longer shopping , yet the finding is somewhat consistent with Kellaris and Kent’s (1992) finding that time does not necessarily fly when exposed to affectively positive musical stimuli. This may have been partially due to misperception differences about how long the task was taking. Shoppers who shopped for a fixed amount of time perceived that they spent more time shopping when listening to familiar music compared to unfamiliar music. In the variable time shopping condition, the greater misperceptions of time in the familiar music compensated for the greater actual shopping time of the unfamiliar music subjects resulting in both groups reporting that they thought they had shopped for an equivalent amount of time. Thus, studies using only reported shopping times to evaluate environmental music may yield inaccurate conclusions.
This complex relationship between music and shopping times demonstrates the practical difficulty of using environmental influences to affect consumer behaviour. An environmental factor may have a predictable and desirable effect on one aspect of shopping, but there is the possibility that it will affect other aspects in unpredictable or undesirable ways. The net result in many cases may be that the desired behavioural change does not occur. Retailers are therefore cautioned to evaluate their atmospheric design choices on all relevant dimensions.
Efforts to relate music to emotional states may also be affected by compensating factors. For example, all three emotional states (Pleasure, Arousal and Dominance) were affected by the interaction of music with control of the amount of time shopped. In the fixed time conditions, listening to unfamiliar music increased perceptions of Pleasure and Dominance but not Arousal. In the variable time condition, unfamiliar music enhanced Arousal but not Pleasure and Dominance. A possible explanation is that in the variable time condition, Pleasure and Dominance may have been enhanced by the unfamiliar music and individuals responded by continuing their shopping (actual shopping was longer in this condition). After the extended time period, emotional states may have reverted back to normal levels such that by the time emotional states were assessed they were equivalent to those in the familiar music condition. If so, this implies that retailers should not rely on post shopping (exit) questionnaires to evaluate emotional states. Instead, interviewing should be done while shoppers are still examining the merchandise.
Although studies have reported significant correlations between measures of retail environments, emotional states and shopping experiences (e.g., Dawson, Bloch and Ridgway 1990; Donovan and Rossiter 1982; Sherman and Smith 1987), the experimental research reported in this paper found the relationships between music, emotions and product evaluations to be complex. Overall, product evaluations were positively related to pleasure. This is consistent with the idea that consumers will be more favourably disposed toward products when they are in a good mood (Gardner 1985). Further, product evaluations were negatively related to arousal. This is consistent with the idea that aroused shoppers may be more vigilant and discriminating when examining products. However, neither emotional measure used in the present research accounted for the effects of familiar music on product evaluations. For shoppers in the fixed time condition, familiar music enhanced evaluations relative to unfamiliar music. This occurred even though pleasure was lower and arousal greater when individuals were exposed to familiar music in the fixed time condition. Clearly, more research is needed to determine if there are other factors besides the emotional states postulated by Mehrabian & Russell (1974) that might cause music to affect product evaluations. One possibility is through classical conditioning in such a way that individuals are not aware of emotional state changes (cf., Gorn 1982). Yet another issue may emphasize the difficulty of interpreting experimental effects of music on shoppers: Kellaris and Kent (1993) warned that, “When different pieces of music are used to manipulate a musical variable, musical properties will be confounded, making it difficult to isolate specific causal antecedents (p. 395).” Nearly all retail studies involving music have used different selections for different experimental conditions.
The study reported in this paper examined environmental music varying in it’s level of familiarity to shoppers. As a complex stimulus, music has many properties that might affect shopping behaviour. Future research should consider aspects such as liking of music (a common factor in advertising experiments on music; e.g., Gorn 1982), tempo (e.g., Milliman 1982), volume (Kellaris & Altsech 1992), and the interaction between music and characteristics of individual listeners (Kellaris and Kent 1993).