Exploring Tempo and Modality Effects, on Consumer Responses to Music

James J. Kellaris, University of Cincinnati
Robert J. Kent, Drexel University
ABSTRACT - Music plays a pervasive role in consumers' lives and is often used in marketing communication. Several sources have recommended "dissecting" music to isolate component properties responsible for various effects on consumers. Recent advances in electronic music technology facilitate such research. Two experiments were conducted to explore tempo and modality effects on listeners' responses to music. Experiment I used computer technology to create music containing orthogonal manipulations of tempo and mode. To assess the generality of effects observed in Experiment I, Experiment II used commercially recorded music to manipulate tempo and mode. A consistent pattern of main and interaction effects was found across the two experiments.
[ to cite ]:
James J. Kellaris and Robert J. Kent (1991) ,"Exploring Tempo and Modality Effects, on Consumer Responses to Music", in NA - Advances in Consumer Research Volume 18, eds. Rebecca H. Holman and Michael R. Solomon, Provo, UT : Association for Consumer Research, Pages: 243-248.

Advances in Consumer Research Volume 18, 1991      Pages 243-248


James J. Kellaris, University of Cincinnati

Robert J. Kent, Drexel University


Music plays a pervasive role in consumers' lives and is often used in marketing communication. Several sources have recommended "dissecting" music to isolate component properties responsible for various effects on consumers. Recent advances in electronic music technology facilitate such research. Two experiments were conducted to explore tempo and modality effects on listeners' responses to music. Experiment I used computer technology to create music containing orthogonal manipulations of tempo and mode. To assess the generality of effects observed in Experiment I, Experiment II used commercially recorded music to manipulate tempo and mode. A consistent pattern of main and interaction effects was found across the two experiments.


Consumers are exposed to music in many contexts, yet relatively little is understood concerning music's effects. Thus it is not surprising to see the recent emergence of interest in music among consumer researchers. Musical studies have recently appeared in consumer esthetics (e.g. Holbrook and Anand 1988), hedonic consumption (e.g. Lasher 1988), mood research (e.g. Bruner 1990), advertising (e.g. Alpert and Alpert 1988; Kellaris and Cox 1989) and retail atmospherics (e.g. Milliman 1982, 1986).

Music is a complex composite of structural elements such as tempo, pitch, and timbre (Bruner 1990), yet it is frequently treated by researchers as a unidimensional stimulus. For example, advertising studies have examined the effects of music's presence or absence (e.g. Park and Young 1986) and the effects of pleasant vs. aversive music (e.g. Gorn 1982). Few studies in our discipline have attempted to isolate the individual components of music responsible for observed effects. If our understanding of music's influence on consumers is to progress, we must recognize the multidimensional nature of music and "dissect" it experimentally.

This approach was advocated by Hevner in the 1930's, but difficulties of stimulus construction were an impediment until recently. Advances in electronic music technology facilitate the construction of musical stimuli necessary for such an experimental approach. For example, music stored as digital information in a computer may be reproduced at various speeds without affecting the pitch. This digital information can be edited to manipulate any dimension of music while keeping all other dimensions constant.

Music has been characterized in terms of three dimensions: time, pitch, and texture (Bruner 1990). The present study explores the effects of two major structural elements of music: tempo and modality. Tempo is an important time-related variable since it controls the pace or "spacing" of sounds. It is a particularly important predictor of human response to music (Hevner 1937; Rigg 1940). Previous studies (e.g. Milliman 1982, 1986) have tended to ignore potential nonlinear and interactive effects of tempo with other musical variables. Modality is an important pitch-related variable since it provides the basic framework within which pitches are organized to form melodies and harmonies. Modality refers to the configuration of intervals between notes in a scale, the most common examples being the "major" and "minor" modes (Apel 1973). Modality is a well-established antecedent of affective response to music (e.g. Heinlein 1928; Hevner 1935a). In general, major keys tend to be-associated with positive thoughts and feelings, and minor keys with negative thoughts and feelings. There are also many "atonal" modalities which are neither major nor minor, each with its own aesthetic character. A recent study by Stout and Leckenby (1988) found mode to have the greatest impact of the musical components examined.

Bruner (1990) proposed that the components of music are capable of producing main and interactive effects on affective, cognitive, and behavioral responses of consumers. He states that "although attempts have been made over the years to understand the main effects of musical components. . . interaction effects have received much less attention" (p. 11). It is important to consider interactive effects since single-variable studies can lead to erroneous conclusions if undetected interactions exist.

The present study tests Bruner's interaction proposition by manipulating two important dimensions of music and measuring multiple outcomes in two experiments. This study extends previous research on tempo and modality effects by exploring nonlinear and interactive effects, and by considering unconventional "atonal" modalities in addition-to major and minor keys. The first experiment used original computer-generated music to provide maximally "clean" manipulations of tempo and modality. Experiment I focuses strongly on internal validity to establish the existence of effects. The second experiment assessed the generality of the effects observed in the first experiment. Experiment II used multiple examples of commercially recorded music to represent orthogonal combinations of tempo and modality.


A 3 X 3 factorial experiment using a between-Ss design was conducted. Treatments included three levels each of musical tempo (fast = 180 BPM, moderate = 120 BPM, slow = 60 BPM) and modality (major, minor, atonal). Outcome measures included multi-item evaluative and behavioral intent scales. The procedure involved randomly assigning Ss to treatment conditions, exposing them to music via headphones, and having them fill out a questionnaire.


One hundred eighty (n = 180) volunteers were recruited from a large introductory class at the University of Cincinnati (U.C.). The sample was 54.2 % male. Ages ranged from 20 to 40 with a median age of 21 years.


The stimuli were audio cassette recordings of original "classical" style instrumental music produced by the authors in a digital sound studio at U.C. Nine versions of one musical composition were produced: one in each of three modes (C major, C minor, and an "atonal" version based on a wholetone scale); with each played at fast (180 BPM), moderate (120 BPM), and slow (60 BPM) speeds. These tempo levels were based on musical convention and precedent. Three levels were chosen because of-the possibility of non-linear effects.

To create the stimuli, one of the authors composed a four-part instrumental fugue in eighteenth century contrapuntal style, based on thematic material taken from an unpublished transcription of a Byzantine chant. The composition was scored for woodwinds and cello, and recorded using digitally sampled sounds. The score was input through an electronic keyboard and stored in digital form using ProPerformer midi sequencer software and a MacIntosh computer. This allowed the authors to edit and output multiple versions of the score representing orthogonal combinations of tempo and mode, while holding all other dimensions of the music constant. The digital information was sent through a polyphonic synthesizer to produce the sound signal.


At the time of recruitment, Ss were told that their participation was sought for a "music study," and offered course credit for participation. As individuals arrived at the lab, they were issued a set of headphones, a randomly assigned numbered audio cassette tape, and a (matching) numbered questionnaire. An attendant directed each S to an audio carrel, where printed instructions guided them through the procedure. After listening to a tape and completing a questionnaire, Ss returned the materials to a lab attendant. Responses were anonymous; Ss signed a separate form to receive course credit. The entire procedure took about twelve minutes.


Dependent variables included multi-item evaluative and behavioral intent scales. Other items were included on the questionnaire but not analyzed in this study.

The evaluative measures consisted of sixteen seven-point semantic differential scales preceded by the prompt "The Music I Heard Was:" The specific items were compiled from Berlyne (1974) and other sources. Half of the items were reverse scored. Two composite scales were formed from these items: the first scale consisted of "appealing, beautiful, likeable, and pleasant" (Cronbach's alpha = .93). We labeled this construct "appealingness." The second scale consisted of "arousing, stimulating, energetic, exciting, and loud" (alpha = .68), which we labeled "arousingness."

The behavioral intent scale consisted of four five-point agreement items relating to intention to listen to the music again if given the opportunity, intention to choose the music over an alternative choice, and intention to purchase the music if it were commercially available. The alpha reliability of the composite scale based on these items was .91.

Other items on the questionnaire included open-ended debriefing items, some standard demographic items, three scales relating to level of interest in music, listening habits, and extent of formal training, and a question on musical preferences.

Results and Discussion

Analyses of variance (ANOVAs) assessed group differences on each outcome measure. (See Table 1.) The analysis revealed significant main effects of mode (F = 36.85; p < .0005) and tempo (P = 9.02; p < .0005), as well as a two-way interactive effect (F = 3.00; p < .02) on appealingness. The magnitude of the interaction was estimated at omega-squared = .03. Figure 1 illustrates this interaction.

The "major" mode was generally evaluated as more appealing than the "minor" mode. This finding is consistent with previous work (Hevner 1935a). As one might expect, the "atonal" versions were the least appealing. This could be due to the greater dissonance, or the unfamiliar sound of this mode.

Consistent with previous work by Holbrook and Anand (1988), the response pattern for the major mode follows an inverted U shape across tempi -- the classic "Wundt curve." The moderate tempo is evaluated as more appealing than slower or faster tempi, even within the fairly restricted range of speeds examined in this study. This is not the case, however, for music pitched in other modes. Both minor and atonal versions were rated as more appealing (or perhaps less unappealing) at faster tempi. The notion of harmonic dissonance (Helmholtz 1862) may explain this finding. The sustained dissonances heard at slower tempi should be perceived as harsher than the relatively briefer dissonances experienced at faster tempi.

An ANOVA on the five-item arousingness scale found only a tenuous positive main effect of tempo (F = 2.73; P < .068), and no other main or interactive effects. Musical tempo, as a psychophysical stimulus property, should be expected to produce greater arousal at higher intensity levels (Berlyne 1974). While this effect is somewhat evident in the data, the range of tempi in this experiment may be too narrow to allow the effect to manifest itself strongly.





Analysis of the behavioral intent scale found significant main effects of both mode (F = 8.71; p < .0005; omega2 = .075) and tempo (F = 5.16; p < .007; omega2 = .04), and a statistically marginal interaction of the two (F = 1.71; p < .15). The effect of tempo on behavioral intent was almost monotonically positive. The modal effect followed a pattern typical of most of the analyses: major keys produced the most positive intent, followed by minor keys, with the atonal versions producing the least positive intent.

To summarize, tempo and modality appear to be important influences on responses to music. Consistent with Bruner's proposition, tempo and modality operated through main effects and interactively. Tempo appears to operate as a psychophysical property of sound -- its influence seems to stem simply from its intensity. Modality effects are more complex. They may stem from the relative consonance/dissonance associated with different modes, or, in the case of atonal modalities, from the relative novelty of the sound.

To explore the generality of the effects observed in Experiment I. a second experiment was conducted using multiple examples of commercially recorded music.


Experiment II is a replication of Experiment I with different musical stimuli. Unfamiliar commercially recorded music was substituted for the computer-generated stimuli used in Experiment I. Three musical selections represented mode and tempo within each cell. The basic design (3 X 3 between-Ss factorial), procedure, and measures remained the same. Composite scale reliabilities (Cronbach's alpha) were .91 and .85 for the appealingness and arousingness scales, and .91 for behavioral intent.


One hundred sixty-two (n = 162) students were recruited for Experiment II. None had participated in Experiment I. Ss were given course credit for their participation and were naive to the purpose of the study. The sample was 53.7% male. Ages ranged from 19 to 35 with a median age of 21 years.

The stimulus materials were audio cassette recordings of instrumental musical excerpts selected to manipulate three levels each of tempo and modality orthogonally. Each excerpt faded to silence after three minutes. Three pieces of music were selected for each cell of the 3 X 3 design to avoid confounding treatment groups with any given piece of music. In each cell, one selection was an obscure classical orchestral piece, one an Indian raga, and one a Chinese piece. The use of obscure classical and non-western instrumental music was intended to avoid problems associated with prior exposure. An attempt was made to minimize confounding treatments with composer, instrumentation, or other musical variables to the extent possible. Of course, using commercial recordings in place of the artificial "perfect" operations in Experiment I is likely to introduce noise in the manipulations; but, the purpose of Experiment II is to see if the effects survive the "imperfect" operations of tempo and modality.

Two important changes were made in the operations of tempo and modality. First, a broader range of tempi was used. Because of the rather weak effect of tempo on arousingness found in Experiment I, tempo levels were set at 60 or fewer BPM (slow), 120 BPM (moderate), and 180 or more BPM (fast). Second, the "atonal" mode in Experiment I was based on a whole-tone scale, which is neither major nor minor. There are numerous other atonal modes: polytonal, serialistic, and minimalistic music are examples from western traditions; non-western music is often based on microtonal systems which are also neither major nor minor. Whereas only one specific type of "atonal" mode was used in Experiment I, a broader range of unconventional tonalities were represented in Experiment II.

Results and Discussion

ANOVA results are summarized in Table 2. The appealingness analysis suggested a pattern of main and interactive effects directionally similar to those in Experiment I; however, the effects were weaker. There are weak main effects of tempo (F = 2.72; p < .069) and mode (F = 2.18; p < .12), as well as a tenuous interactive effect (F = 1.94; p < .106). (See Figure 2.)

The inverted U-shaped 'Wundt curve" remains evident among the major mode groups. The positive effect of tempo on evaluations of atonal music is also evident.

An ANOVA on the arousingness scale revealed significant main effects of both tempo (F = 22.09; p < .0005) and modality (F = 2.97; p < .05). As in Experiment I, no interactive effect was found. The effect of tempo was positive, i.e., faster music was evaluated as more arousing. The modality effect was slightly different from that observed in Experiment I. In both experiments music pitched in major keys was evaluated as least arousing and atonal music was more arousing; however, in Experiment II minor modalities produced higher arousal ratings than atonal music.

Finally, the ANOVA on behavioral intent found significant main effects of tempo (F = 3.19; p < .044; omega2 = .025) and modality (F = 3.99; p < .021; omega2 = .034), and a tenuous tempo-mode interaction (F = 2.01; p < .095). The statistical results are similar to those found in Experiment I. Again, slower tempos and atonal modalities produced the least positive behavioral intent.

In sum, the pattern of main and interactive effects of tempo and modality in Experiment II was very consistent with Experiment I. In the case of appealingness ratings, the tempo-mode interaction effect was weaker in-Experiment II. This is what one might expect given the less pure (but more realistic) manipulations in Experiment II. The arousal effect of tempo was more pronounced given the expanded range of tempi represented in Experiment II. The effects on behavioral intent were fairly consistent across the experiments, with atonal music producing the least positive intent.


This study has explored the effects of two important objective stimulus properties of music (tempo and modality) on consumers' responses to music. Findings were generally supportive of Bruner's (1990) proposition concerning interactive effects of music's components. A consistent pattern of effects was found across two experiments using different operations. Tempo was found to have positive main effects on evaluations of music's arousingness and on behavioral intent. Modality also influenced arousal and intent, with atonal modalities producing the least positive responses. Tempo and mode were found to operate interactively on evaluations of music's appealingness, and, to some extent, on behavioral intent.





Some limitations of the study should be recognized. First, the instrumental music examined in this study was drawn from classical sources. Thus, the findings might be more characteristic of "art music" as opposed to popular/commercial forms of music. Second, the duration of exposure to the musical stimuli was relatively brief (three minutes). It is possible that different responses to music could develop after more prolonged exposures, such as in a store, restaurant, or concert setting. Third, the sample has both advantages and disadvantages. The use of student subjects offers statistical advantages of homogeneity, but may impose limitations on generality. College students spend more time listening to music than the general adult population, and their musical tastes lean strongly toward current Rock/Pop. (68.5% of the sample across the two experiments reported preference for this listening category.) On the other hand, this may be an advantage of the sample because treatment effects of classical and non-western music would be less likely to show themselves among Ss who generally lack knowledge and interest in these musical genres. The use of less familiar, less preferred music constitutes a stronger test for treatment effects.

This study yields a number of preliminary implications. First, it is clear that music's structural components (at least tempo and mode) can work interactively as influences on listeners. This implies that individual components of music should not be treated in isolation in research or practical applications. Second, music pitched in major keys seems to generate the most positive responses; however, given the 'Wundt curve" effect, tempi which are either too fast or slow are likely to mitigate this influence. Marketers or other parties wishing to take advantage of music's positive influences should search for optimal tempi at which hedonic response is maximized. Third, under certain circumstances it may be desirable to use minor or atonal music because it provides variety or is more aesthetically congruent with an advertised product or message (e.g. dissonant music in anti-drug abuse ads). In such cases, faster tempi might mitigate negative thoughts and feelings, and slower tempi might intensify negative effects. Finally, the findings on behavioral intent toward music may generalize to behavioral intent toward products via an associative process when music is used in ads or in retail environments. This could be explored in future research.

Several other directions for future research are suggested by this study. For example, effects of other musical components such as sound texture could be examined. Objective stimulus properties of music such as tempo and modality might operate through "subjective" stimulus properties such as novelty or complexity (Berlyne 1974). It would be interesting to explore the mediating role of such variables, and perhaps to assess the relative predictive power of objective vs. subjective stimulus properties of music. Finally, this research could be extended by studying the effects of musical components on the processing of verbal material (e.g. song lyrics, advertising messages). The aesthetic congruency of music with words may influence the reception and interpretation of verbal material.


Alpert, Judy I. and Mark I. Alpert (1988), "Background Music as an Influence in Consumer Mood and Advertising Responses," in Thomas K. Srull, ed., Advances in Consumer Research, Vol. 16, Association for Consumer Research, Provo, UT, 485491.

Bruner, Gordon C., III (1990), "Music, Mood, and Marketing," Journal of Marketing, 54, 4 (October), 94-104.

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

Hevner, Kate (1935a), 'The Affective Character of the Major and Minor Modes in Music," American Journal of Psychology, 47, 103-118.

Hevner, Kate (1935b), "Expression in Music: A Discussion of Experimental Studies and Theories," Psychological Review, 42, 186-204.

Hevner, Kate (1937), 'The Affective Value of Pitch and Tempo in Music," American Journal of Psychology, 49, 621-630.

Holbrook, Morris B. and Punam Anand (1988), "Aims, Concepts, and Methods in Marketing Research on Consumer Esthetics: The Effects of Tempo on Perceptual and Affective Responses to Music," unpublished working paper, Columbia University.

Kellaris, James J. and Anthony D. Cox (1989), "The Effects of Background Music in Advertising: A Reassessment," Journal of Consumer Research, 16 (June), 113-118.

Lasher, Kathleen T. (1989), "Hedonic Consumption: Music as a Product," in Thomas K. Srull, ed., Advances in Consumer Research, Vol. 16, Association for Consumer Research, Provo, UT, 367-373.

Milliman, Ronald E. (1982), "Using Background Music to Affect the Behavior of Supermarket Shoppers," Journal of Marketing, 46 (Summer), 86-91.

Milliman, Ronald E. (1986), "The Influence of Background Music on the Behavior of Restaurant Patrons," Journal of Consumer Research, 13 (September), 286-289.

Park, C. Whan and S. Mark Young (1986), "Consumer Response to Television Commercials: The Impact of Involvement and Background Music on Brand Attitude Formation," Journal of Marketing Research, 23(February), 11-24.

Stout, Patricia A. and John D. Leckenby (1988), "Let the Music Play: Music as a Nonverbal Element in Television Commercials," in S. Hecker and D.W. Stewart, eds., Nonverbal Communication in Advertising, Lexington, MA, D.C. Heath, 207-223.

A complete reference list is available from the first author on request.