Calling the tune : how verbal cues referring to producer, repertoire title, and artist on music artwork covers raise song evaluation

Calling the tune: How verbal cues referring to producer, repertoire title,

and artist on music artwork covers raise song evaluation

Elaine Ching 10084738 Master thesis

Graduate School of Communication

Master’s Programme Communication Science: Persuasive Supervisor: E. De Waal

Abstract

A pioneering research based on previous literature regarding the use of verbal cues on products and advertisements, and inspired by similar use in movie marketing and proven effective when referring to a previous box-office movie by the same director (e.g. “From the director of Avatar”), the use of verbal cues tailored for music products such as referring to producer, previous repertoire title and artist name on music artwork covers was explored, and whether familiarity with verbal cues enhances song evaluation. The study consisted of a one-factor experimental design with additional verbal cues as the one one-factor with six different levels displaying either a singular or combination of cues (e.g. “Produced by [producer] who produced [“repertoire title]”). A total of 312 Dutch participants took part in the study and the results confirm that the use of additional verbal cues was more effective in raising song evaluation than the absence of cues, and did not interact with familiarity. In fact, familiarity with additional verbal cues has a direct effect on raising song evaluation, and in particular, familiarity with repertoire title and artist name has shown to be most effective in raising song evaluation, whereas familiarity with producer did not. However, considering that verbal cues can consist of factual information (i.e. reference to product attribute) or familiar names (i.e. reference to brand name), it is the combination of all three additional cues: “Produced by [producer] who produced [repertoire title] by [artist name]”, that seems to be the optimal combination of additional verbal cues based on familiarity with repertoire title and artist name, and the factual information provided by who has produced the song.

Introduction

A music artwork cover is a visual representation of a single piece or collection of music, by one or more individual artists or a collective band of artists. The artwork cover is a crucial element in music organising (Schindler, 2014), selecting (Bainbridge, Cunningham &

Downie, 2004) and discovering new music (Libeks & Turnbull 2011). In fact, as outlined in a seminal paper regarding artwork covers by Jones and Sorger (1999), in most cases the music artwork cover is the foundation of a marketing campaign from which other marketing assets are developed such as the music video, merchandise, tour posters and advertisements. More importantly, the target group for the artwork cover and its subsequent marketing assets are not merely consumers only. The group includes professionals from different areas of the music industry such as radio DJs, playlist curators and music journalists, whose judgement can be decisive in further reach and success. Furthermore, the artwork cover can be the determining factor that can convince target groups to play a new record or ignore it all together, and in case of the latter, the subsequent marketing assets can possibly equally fail. For example, a consumer who does not listen to a record will subsequently be least likely to buy tickets for a tour. Taking these points into consideration, it has become apparent that the design and use of a music artwork cover should no longer be considered as solely a piece of expressive art. Instead, artwork covers should be explored and improved upon in order to anticipate its future strategic use and importance in music marketing.

An illustration of a possible strategic use of an artwork cover is best described when drawing on examples from neighbouring entertainment industries such as the film industry. When introducing new movies to the general public, similar marketing assets such as posters, trailers, and advertisements are developed and distributed (Karray & Debernitz, 2017). Aside from movie stills or images of the actors, the movie title and the names of the actors are clearly displayed. These elements are similar to the mandatory elements of a music artwork

cover in order for a graphic design to constitute as a music artwork cover (e.g. graphic, song title, and artist name). However, a common practice within the film industry is to mention who has directed the movie and in some cases, when the actors are relatively unknown, to name the director in conjunction with previously directed films (Finsterwalder,

Kuppelwieser, & De Villiers, 2012). The reasoning behind this practice assumes that a

director who has achieved box-office successes with previous movies can function as a cue to influence positive attitudes towards a new film (i.e. movie evaluation), due to familiarity with displayed cue (D’Astous & Touil, 1999). This is usually established by adding verbal cues such as “From the director of Avatar” or “Directed by James Cameron who directed ‘Avatar’”, which indeed has been proven effective in raising positive attitudes towards the new movie prior to the movie experience: consumers who recognize the mentioned movie will be more inclined to associate their positive associations with a positive expectation of the new movie (Sauer, 2014). A translation of this strategic use in movie marketing to music marketing would suggest exploring the use of similar additional verbal cues on music artwork covers in order to raise positive attitudes towards respective new song (i.e. song evaluation). Furthermore, whether its effect is similarly amplified due to a degree of recognition or familiarity with additional verbal cues. For example, by mentioning the producer of the song or in combination with a previously produced repertoire song title on music advertisements (e.g. “Produced by Quincy Jones”, or “Produced by Quincy Jones who produced ‘Billie Jean’”). However, the pairing between producer and previous repertoire song title is

incomplete without mentioning the artist of the previous repertoire song title (e.g. “Produced by Quincy Jones who produced ‘Billie Jean’ by Michael Jackson”). In fact, it is more

common to find the pairing between repertoire song title and artist, as one cannot go without the other when new music is released. Hence, when adding cues to the music artwork cover such as referring to the (1) producer of the song, (2) previously produced repertoire song title

(3) the respective artist of the repertoire song title mentioned, this adds three additional verbal cues on a music artwork cover aside from the verbal cues musician name and song title.

The initial concept and use of verbal cues have been researched as verbal snippets of information which are placed on advertisements, physical products or packaging, with the intent of assisting the consumer to retrieve prior knowledge and associations that can raise brand or product evaluations (Keller, 1987, 1991). Hence, in earlier research, the verbal cue has been explored as an additional leverage aside from brand name and physical product, to persuade consumers to prefer one brand or product over another, or associate one brand with another (Aaker & Keller, 1990; Janiszewski, Kwee, & Meyvis, 2002; Janiszewski & Van Osselaer, 2000; Van Osselaer & Janiszewski, 2001). However, additional studies have shown that the effect on product or brand evaluation is either dependent or further influenced by familiarity with cues such as well-known brand names (Baker, Hutchinson, Moore, &

Nedungadi, 1986; McDaniel & Einstein, 1993; Rindfleisch & Inman, 1998; Simonin & Ruth, 1998). Moreover, familiarity influences the degree of recognition, as cues that are familiar increase levels of recognition opposed to cues that are unfamiliar, and thus, within studies focused on verbal cues, the role of recognition is inherent to familiarity (Brennan & Babin, 2004). Thus, suggesting, that aside from verbal cues having a direct effect on evaluation, the effect can be amplified when verbal cues are indeed deemed familiar. In fact, studies

conducted in the neighbouring film industry has shown that familiarity with star power (i.e. director, actors, etc.) are key factors in raising positive movie evaluation and success (D’Astous & Touil, 1999; Desai & Basuroy, 2005; Finsterwalder et al., 2012; Litman & Kohl, 1989).

Yet, the influence of verbal cues or star power wrapped as verbal cues, in music marketing has never been researched. The previously mentioned paper by Jones and Sorger (1999) was published nearly two decades ago and the authors demonstrated their foresight

when they called upon further research in the development of artwork covers. Urging scientists and marketers alike to improve our understanding of artworks covers within the fields of (music) marketing. They believed that the artwork cover would remain important in the new era of streaming. However, surprisingly, in the past twenty years, no scientific research regarding verbal cues on music artwork covers from a marketing perspective has been conducted. This has created a looming gap in knowledge between the end of the century and current times, and it is alarming to discover that the music industry and in particular music marketing, has been overlooked in scientific research. Particularly, in comparison with the movie marketing industry wherein ample studies with regards to the use of verbal cues has been explored in whether it can raise movie evaluation and predict success. Therefore, in an attempt to lay a foundation for future research and build a bridge between scientific research and practical usage, this study will contribute by providing a new perspective and approach to music marketing, and in specific: the strategic use of additional verbal cues on an artwork cover in raising song evaluation. Furthermore, whether the effect of additional verbal cues is amplified by familiarity with cues. This leads to the formulation of the following research question;

RQ: To what extent do additional verbal cues (producer, previously produced repertoire title and artist name) on a music artwork cover impact song evaluation and to what extent is this amplified by familiarity with additional verbal cues?

Theoretical Framework Presence of additional verbal cues

The pairing of (verbal) cues with a brand or product and its effect on raising evaluation of target brand or product, has been extensively researched based on the assumption that this

pairing forges new associations with prior knowledge which allows for positive product or brand evaluation (Keller, 1987, 1991) and prediction of benefits from using product or brand (Van Osselaer & Janiszewski, 2001). However, before diving further into the effects of additional verbal cues, the first question is whether the use of additional verbal cues opposed to the absence of verbal cues has an effect on raising song evaluation. According to the findings of a study conducted by Keller (1987), the addition of cues in advertisements increased positive product and brand evaluation opposed to the absence of cues (Keller, 1991). Similarly, previous studies focused on the effects of additional cues in movie marketing have shown that the inclusion of cues raises positive evaluation of target movie than when cues are omitted (Elberse, 2007; Finsterwalder et al., 2012). Furthermore, the breadth of research on the effects of verbal cues raising product or brand evaluation is divided between the use of verbal cues containing factual information (Abernethy & Franke, 1996; Keller, 1987) and verbal cues referring to other known brand names (Aaker & Keller, 1990; Janiszewski & Van Osselaer, 2000). The main difference between these studies is that the use of verbal cues containing brand names require a degree of familiarity, whereas verbal cues containing factual information does not. Hence, depending on which type and content of verbal cue, familiarity may not be mandatory to effectively raise evaluation. In fact,

according to the cue compatibility principle as formulated by Keller (1991), in order to achieve a positive brand or product evaluation, the factual information of the cue must match the associations retrieved from memory in relation to the new product or brand. In other words, the cue must be directly relevant in retrieving associations that can indeed assist in predicting the benefit of a new product or brand, which in turn can raise evaluation (Dawar & Anderson, 1994; Meyvis & Janiszewski, 2002). For example, the phrase “Produced by

[producer] known for [repertoire title] by [artist]” informs consumers about the production of the new song, suggesting that it is a key component or ingredient of the song, which in itself

can raise evaluation even if the consumer is not familiar with either of the three additional verbal cues. Hence, additional verbal cues on music artwork covers referring to the producer, previously produced repertoire title and respective artist are directly relevant as they provide factual information with which the new song can be evaluated regardless of familiarity with cues (Carpenter, Glazer, & Nakamoto, 1994). In fact, according to Van Osselaer and

Janiszewski (2001), a feature, such as an ingredient of a product, can function as a cue with which consumers can learn to predict benefits of a brand or product regardless of familiarity. Which is exactly the intent of a verbal cue on an artwork cover: these additional cues can assist consumers in predicting the benefit of listening to a new song which in turn can raise song evaluation. Thus, based on these findings and given that music artwork covers have yet been explored with the addition of verbal cues compared to the absence of additional verbal cues, and that these cues contain factual information which does not necessarily require familiarity, the following hypothesis will be researched as is displayed in figure 1:

H1: Additional verbal cues on artwork covers will lead to more positive song evaluation, than the absence of additional verbal cues on artwork covers.

Familiarity with additional verbal cues

The additional verbal cues on music artwork covers referring to either producer, previously produced repertoire title and respective artist can provide factual information regarding the new song regardless of familiarity. However, they are in itself brand names that can be deemed familiar. In the breadth of research regarding verbal cues containing familiar brand names, there are multiple different threads such as spillover studies (Balachander & Ghose, 2003; Simonin & Ruth, 1998), brand alliance studies (Janiszewski et al., 2002; Rodrigue & Biswas, 2004) and co-branding studies (Baumgarth, 2004; Washburn, 2000), that all weave

into the basic premise wherein associations with familiar brands or products can help raise the evaluation of unfamiliar brands or products (Aaker & Keller, 1990; Sheinin, 2000; Swaminathan, Reddy, & Loughran Dommer, 2012). Furthermore, according to studies focused on (verbal) cues in movie marketing, the degree of familiarity with cues are considered key factors in raising positive movie evaluation, and in particular for directors (D’Astous & Touil, 1999; Simonton, 2009), previously directed movies (Sauer, 2014) and actors (Elberse, 2007; Finsterwalder et al., 2012; Litman & Kohl, 1989). Therefore, moving onto the next question beyond the absence or presence of cues is whether being familiar with additional verbal cues (i.e. producer, previously produced repertoire title and respective artist) raises the evaluation of the unfamiliar verbal cues (i.e. new song). First, according to Janiszewski and Van Osselaer (2000), the use of verbal cues containing familiar brand names can help predict the benefit and performance of target products. Second, according to the study results by Simonin and Ruth (1998), a brand alliance wherein the use of verbal cues containing familiar brand names are paired with an unfamiliar brand name can raise positive evaluation of target product. Furthermore, their study results suggest that this relation is moderated by the degree of brand familiarity with additional verbal cues which strengthens the positive evaluation of the target product (Rodrigue & Biswas, 2004).

Thus, based on these findings and given that music artwork covers have yet been explored with a variety of individual verbal cues (i.e. producer, previously produced repertoire title and respective artist) and that the degree of familiarity with each cue can amplify the effect on song evaluation, the following hypothesis will be researched as is displayed in figure 1:

H2: The effect of additional verbal cues on music artwork covers on song evaluation will be greater when the cues are familiar than when not familiar.

Combination of different additional verbal cues

To expand this study and explore beyond the absence versus presence of additional verbal cues, and familiarity, a closer look will be given to the number of additional verbal cues. While the previous hypotheses consider all additional verbal cues, a closer observation should be given to the cumulative effect of one to three cues. A review of the literature appears to be divided between studies suggesting to use a singular cue wherein one side argues that referring to multiple brands lowers brand or product evaluation (Keller, 1987; Meyvis & Janiszewski, 2002), and the other side suggesting that a combination of two or more cues can raise evaluation (Laroche, Kim, & Zhou, 1996; Miyazaki, Grewal, &

Goodstein, 2005). Furthermore, in support of studies suggesting multiple cues, studies on the use of cues in movie marketing suggest that the combination of cues and in particular the familiarity with each cue, could collectively raise movie evaluation, although one cue can be more effective than the other (Elberse, 2007; Finsterwalder et al., 2012). However, no earlier study has, in fact, explored the cumulative effect of additional cues in movie marketing and thus, a translation to music marketing allows for a testing of these assumptions. Hence, to expand the current study the following assumptions will be explored:

Assumption A: More additional verbal cues effectively raise song evaluation with each added verbal cue.

Assumption B: More familiar additional verbal cues effectively raise song evaluation with each added familiar verbal cue.

Based on both assumptions, it raises the question whether the combination of more additional verbal cues and more familiarity with each cue raises song evaluation. Hence, the following sub-question will be researched:

SQ: Does more additional verbal cues and more familiarity with each cue effectively and respectively raise song evaluation?

The additional exploration of these assumptions and sub-question allows for this study to expand beyond the body of research regarding verbal cues whether absent or present and familiar or unfamiliar, and in specific, its effect on song evaluation. Moreover, it can provide insight into the use of additional verbal cues beyond current studies regarding verbal cues and in movie marketing. Thus, the exploration of these assumptions and sub-question allows for an advantage with regards to how many cues, how many familiar cues, and how the

combined effect of multiple and familiar cues can raise song evaluation.

Figure 1. Conceptual model

Methods Design and participants

The study was a between-subjects one-factor design with one independent categorical factor named ‘additional verbal cues’ consisting of six levels each representing a different type or

Additional verbal cues: 1. None

2. Producer 3. Repertoire title

4. Repertoire title and artist name 5. Producer and repertoire title

6. All cues (producer, repertoire title and artist name)

Song evaluation Attitude towards new song of unknown musician Familiarity with

additional verbal cues: 0 = not familiar 1 = familiar

H1+ H2+

combination of cues. In December 2017, an online experimental survey in Qualtrics was widely distributed via social media and reached 450 Dutch participants who consented to partake in the study. All participants were randomly sorted into one of the six respective conditions, as shown in table 1. A total of 138 cases were excluded from analysis due to leaving the study prematurely, resulting in missing values. The final sample size consisted of 312 participants of which 69.7% were female and 32.1% male. The majority of participants completed a higher educational degree (58%) and did not work in the music industry

(81.4%). The mean age of all participants was 27.4 years (SD = 10.10), ranging from 16 to 61 years of age.

To confirm a successful manipulation of the unknown musician, song title, and music genre, a frequency table was run to verify that all participants are unfamiliar with fictitious musician and song title, and the majority of participants were unfamiliar with the musician (98.7%, n = 308) and song title (99%, n = 309). A second frequency table indicated that an overwhelming majority of participants identified the music genre of the music artwork cover belonging to urban music sub-genres, in particular, rap (51%, n =159), hip-hop (38.8%, n =121), R&B (5.1%, n =16) and soul (.3%, n =1). The remaining 4.8% (n = 15) identified other genres or expressed indifference.

Materials

The visual stimuli consisted of six different versions of the music artwork cover. The music artwork cover is designed to belong in the urban music genre which consists of, but is not limited to, the sub-genres R&B, rap, hip-hop, soul, and funk. The music artwork cover displayed a red background with a graphic illustration of a male silhouette wearing a baseball cap, sunglasses, gold chains, and smoking a cigarette. In the upper left corner, the name “Vincent Oscar” was placed, which is a fictitious name and an equally unknown Dutch song

title “Raak” was placed on the right of the illustration. The first version (1) “none” consisted of no other visual or verbal cues placed on the cover. The other five conditions displayed additional cues that were centred at the bottom of the artwork cover and consisted of either a singular cue or a combination of cues according to a standard phrase wherein either all cues were displayed or rephrased according to the specific condition. For example, (2) producer name: “Produced by [producer]”; (3) repertoire title: “Produced by producer of [repertoire title]”; (4) repertoire title and respective artist name: “Produced by producer of [repertoire title] by [artist name]”; (5) producer name and repertoire title: “Produced by [producer], who produced [repertoire title]”; and lastly, a combination of all three cues (6): “Produced by [producer] who produced [repertoire title] by [artist name]”. See Appendix A for an overview of all six visual stimuli.

The chosen additional verbal cues refer to an existing producer named Esko who produced the number one hit single “Ride or Die” performed by rapper Josylvio who has released multiple singles and albums. It was the first number one single by this artist and producer, and remained number one for exactly one week in the Dutch charts in September 2017. After which, the single remained in the top-50 chart for several weeks on varying positions. Choosing this particular song controlled the cues from being overly familiar (i.e. number 1 for several weeks) or too unfamiliar (i.e. never having reached number 1).

Procedure

The online experimental survey commenced with a request for consent of participants who were sixteen years of age or older. A brief and general description of a music artwork cover was provided after which through randomization one of the six manipulated artwork covers was shown. Immediately following the exposure of the artwork cover, the dependent variable song evaluation was questioned. After which, questions regarding the manipulation check

and familiarity questions were asked specifically about the condition presented. Lastly, the demographic variables and personal music genre preferences regarding urban music were questioned. After filling in all questions, participants were given the opportunity to enter comments about the online experiment or skip to successfully complete the survey.

Measures

Manipulation check. To accurately assess whether participants correctly observed their respective condition, a multiple-choice question was presented with six answer-items representing all possible verbal cues, including additional verbal cues: (1) musician name (2) song title; (3) reference to producer without mentioning producer name; (4) reference to producer with mentioning producer name; (5) reference to previously produced repertoire title by producer without mentioning artist name; and lastly, (6) reference to previously produced repertoire title with mentioning artist name. The accompanying question, “Please state which text elements were displayed on the music artwork cover”, allowed participants to choose multiple answer-items as their answers based on their respective condition.

Song evaluation. The variable song evaluation was measured on a 7-point Likert scale ranging from 1 (highly unlikely) to 7 (highly likely) based on 4 statements. Two statements measured potential success: “I think this will become a hit record (e.g. ranking high on a music chart)” and “I think this song has a great chance of becoming a hit record”. The remaining two statements measured quality: “The music production appears to be of high quality” and “The overall quality of this record appears to be of high quality”. A principal component analysis showed that all these items load on one factor with an eigenvalue above 1 and proved to be a reliable scale (EV = 2.58, R² = 64.6, ⍺ = .82). The mean scores for the song evaluation scale (M = 3.83, SD = 1.11), are used in analyses wherein a higher mean score indicates a greater and positive song evaluation.

Recall. Two successive recall questions were asked to accurately measure whether participants recalled the additional verbal cues according to each respective condition except for the condition without any additional verbal cue. The first question asked to enter the name of the additional verbal cue (e.g. “What was the name of the producer mentioned on the artwork cover?”) and if not directly recalled, a second question assisted recall by asking to choose the correct cue from a list of five answer-item options (e.g. “Do you recall the producer mentioned on the artwork cover from the list of producers below?”). Per additional verbal cue (i.e. producer, repertoire title and artist) these two question items were computed to one variable ‘recall’ wherein 1 represents participants who either directly recalled the additional verbal cue or with assistance, and 0 for those who did not even after assistance. An overview of participants correctly recalling additional verbal cues per condition is shown in table 1.

Familiarity. While it is possible that incorrect recall could have happened due to a variety of reasons such as distraction or lack of short-term memory, this does not imply unfamiliarity. Thus, regardless of whether the additional verbal cues were correctly recalled, participants were asked if they were familiar with additional verbal cue (e.g. “The mentioned producer was Esko. Are you familiar with this producer?”), wherein 1 represents familiar and 0 for unfamiliar. An overview of participants who are indeed familiar with additional verbal cues per condition, regardless of correct recall, is shown in table 1.

Demographics. Four questions in total were asked with regards to demographics which were age entered in digits, an ordinal multiple-choice question regarding education ranging from lower education to higher education, and two binary questions regarding gender (i.e. male or female) and whether they had an occupation within the music industry (i.e. yes or no; e.g. record labels, radio, etc.).

Table 1

Additional verbal cues conditions: Cues recalled and recognised

Producer cue Repertoire cue Artist name cue Additional verbal cues n Rcll. Fam. Rcll. Fam. Rcll. Fam.

1. None 52 n/a n/a n/a n/a n/a n/a

2. Producer name 57 43 13

3. Repertoire title 53 43 21

4. Repertoire title and artist name

42 29 16 27 14

5. Producer name and repertoire title

53 36 11 37 16

6. All cues (producer, repertoire title and artist name)

55 37 19 34 21 31 18

Total 312 116 43 143 74 58 32

Note. N = 312, all scores are count, Rcll. = recalled; Fam. = familiar.

Control measures

Design evaluation. Given that artwork covers are a visual stimulus which displays design elements other than verbal cues, these visual elements can subconsciously influence song evaluation. Thus, to accurately isolate the effects of verbal cues from the visual cues and to avoid confounding results, the overall attitude towards the aesthetics of the music artwork cover need to be controlled for in further analyses. The question “What do you think of the design of the artwork cover?” asked participants to judge six polarised scales based on a 7-point Likert scale: “old fashioned-modern”, “ugly-pretty”, “stupid-fun”, “unclear-clear”, “unattractive-attractive” and lastly, “low quality-high quality”. A principal component

analysis showed that these items load on one factor with an eigenvalue above 1 and proved to be a reliable scale (EV = 3.33, R² = 55.6, ⍺ = .84). The mean scores for design evaluation

(M= 4.04, SD = 1.17), are to be considered as covariates in analyses wherein a higher mean score indicates a greater and positive attitude towards the overall visual design of the music artwork cover.

Urban music preference. To accurately isolate the effects of verbal cues from the listening preference to urban music, which in turn could also have an effect on facilitating familiarity due to frequent exposure to respective cues, these preferences need to be

controlled for to avoid confounding results. Five items represented sub-music genres of urban music (i.e. R&B, rap, hip-hop, soul, and funk) and are each measured on a 7-point Likert scale from 1 (“Do not like to listen at all”) to 7 (“Like to listen a lot”). A principal component analysis showed that these items load on one factor with an eigenvalue above 1 and proved to be a reliable scale (EV = 2.49, R² = 49.9, ⍺ = .73). These items are computed to average mean scores (M= 4.52, SD = 1.25), wherein a higher mean score indicates a greater preference for urban music and this variable will be added to analyses as a covariate.

Results Randomisation

To ensure comparable representations of participants’ age, gender, level of education, design evaluation, urban music preference, and occupation in the music industry across the six additional verbal cue conditions, a series of One-Way ANOVA, Pearson’s correlation and Chi-square analyses were conducted to examine distributions. A One-way ANOVA with age as the dependent variable and additional verbal cues conditions as the independent variable showed that the mean age across all cue conditions did not significantly differ (F(5, 306) = .75, p = .586). Furthermore, A Chi-square test examined gender, education, occupation in the music industry, and additional verbal cue conditions as categorical variables and showed an insignificant relationship between cue conditions and gender (χ2(5) = 4.48, p = .483),

education (χ2(20) = 27.62, p = .119), and occupation in the music industry (χ2(5) = 2.69, p = .747). A One-Way ANOVA showed that there was no significant relationship between control variables design evaluation (F(5, 306) = 1.18, p = .319), urban music preference (F(5, 306) = .68, p = .828), and additional verbal cues conditions as the independent variable. All variables appear to be well distributed across all conditions.

However, as a second measure to confirm that these control variables do not need to be controlled for as covariates in relation to the dependent variable song evaluation and moderating variable familiarity, a Pearson’s correlation analysis was conducted as shown in table 2. The test showed that there was a significant moderately strong positive relation between dependent variable song evaluation and design evaluation (r = .45, p < .000), a weak positive relation with urban music preference (r = .18, p = .002) and a weak positive relation with gender (r = .14, p = .017). In addition, there was a significant weak negative relation between moderating variable familiarity and urban music preference (r = -.22, p < .000) and a weak positive relation with occupation in music industry (r = .17, p = .006). Thus, the four control variables gender, occupation in the music industry, design evaluation and urban music preference were considered as covariates in analyses to avoid confounding results.

Table 2

Correlations between control variables and song evaluation, and familiarity

Control variables Song evaluation Familiarity

Age .02 .08

Gender .14* -.06

Education -.02 .11

Occupation in music industry .03 .17**

Design evaluation .45** .04

Urban music preference .18** -.22**

Manipulation check

To confirm a successful manipulation between the six additional verbal cue conditions and correctly identifying the cues shown within each condition, a Chi-square test was conducted with the categorical variables ‘additional verbal cues’ and ‘manipulation check’. The name of the musician (χ2(5) = 3.36, p = .65) and song title (χ2(5) = 10.69, p = .058) is insignificantly correlated across all conditions. For the conditions ‘producer’, ‘producer and repertoire title’ and ‘all cues’, and the manipulation check ‘producer cue mentioning producer name’, there is no significant correlation found (χ2(2) = 1.11, p = .058). For the conditions ‘repertoire title’ and ‘producer and repertoire title’, and the manipulation check ‘repertoire title cue without mentioning artist’, no significant relationship was found (χ2(1) = 1.51, p = .22). Lastly, the conditions ‘repertoire title and artist name’ and ‘all cues’, and manipulation check ‘repertoire title cue with mentioning of artist’, showed no significant correlation (χ2(1) = 1.51, p = .22). Across all conditions and cues, there was no significant correlation found in correctly identifying which cues have been shown in each condition. Aside from assuming a general lack of correlation, another possible explanation could be that the wording of the

manipulation check question and answer items were confusing in wording or simply overwhelming to remember correctly, as shown in Appendix B.

Therefore, in another attempt, a Chi-square test was conducted with the variables ‘recall’ of individual additional verbal cues displayed (i.e. producer, repertoire title, and artist name) and ‘additional verbal cues’, based on the assumption that recall can be a substitute for the manipulation check question across the six cue conditions given that recall suggests a manipulation between conditions. The correlation between correctly recalling the name of the producer and conditions displaying the producer name was found to be insignificant (χ2(2) = 2.37, p = .306), as was for recalling repertoire title and conditions displaying repertoire title (χ2(3) = 6.08, p = .108), and recalling artist name and conditions displaying artist name (χ2(1)

= .62, p = .430). Hence, after having exhausted all possible attempts to confirm a successful manipulation between the six additional verbal cue conditions, these results confirmed that in order to proceed with further analyses and hypotheses testing, that instead of continuing with the initial independent variable ‘additional verbal cues’, the variable ‘recall’ (of individual additional verbal cues: producer, repertoire title and artist name) should be considered instead as a reliable substitute, as it can be assumed that participants have seen a different additional verbal cue due to correct recall.

Analyses

The first hypothesis assumes that the addition of verbal cues, regardless of which specific cue and familiarity, will have a greater and positive effect on song evaluation than no additional verbal cues. The substitute independent variable ‘recall’ was not applicable for the cue condition that displayed no cues. Thus, as a first and only exception in this study, an ANCOVA analysis was conducted with the originally intended independent variable

‘additional verbal cues’ with the six conditions sorted between two groups named ‘conditions with cues’ and ‘conditions without cues’, and song evaluation as the dependent variable ranging from 1 (highly unlikely) to 7 (highly likely). The variables gender, design evaluation, and urban music preference are controlled for as covariates and design evaluation proved significantly correlated F(1,306) = 73.65, p < .000, ηp2= .19, as was urban music preference, F(1,306) = 7.52, p = .006, ηp2 = .02, whereas gender was not, as shown in table 3. After

controlling for the effects of covariates, there was a small significant main effect found between conditions with and without cues and song evaluation, F(1, 306) = 4.09, p = .044, ηp2 = .01. Participants who were exposed to a music artwork cover containing cues regardless

of which cue and familiarity, showed a significant and more positive evaluation of the song (M = 3.86, SD = 1.13) than participants who viewed no cues (M = 3.67, SD = .99), as shown

in table 3. Thus, hypothesis one is accepted: cues on artwork covers regardless of which type of cue and familiarity has a positive effect on song evaluation.

Table 3

ANCOVA analysis: Additional verbal cues conditions grouped between with and without additional verbal cues and dependent variable song evaluation

Main effects df F p ηp2 MwithCues MwithoutCues

Additional verbal cues conditions

1, 306 4.09 .044 .01 3.86 3.67

Covariates

Design evaluation 1, 306 73.65 .000 .19 Urban music preference 1, 306 7.52 .006 .02

Gender 1, 306 3.59 .059 .01

Note. N = 312.

The second hypothesis assumes that the effect of additional verbal cues on artwork covers on song evaluation will be greater when the cues are familiar than when unfamiliar. An ANCOVA analysis was conducted with ‘recall’ as the substitute independent variable sorted between those who indeed correctly recalled either additional verbal cue (i.e. producer, repertoire title and artist) and those who did not, and song evaluation as the dependent

variable. Familiarity was considered as the moderating variable sorted between those who are indeed familiar with either additional verbal cue and those who are unfamiliar. The variables gender, occupation in music industry, design evaluation, and urban music preference were controlled for as covariates and design evaluation proved significantly correlated F(1,252) = 61.07, p < .000, ηp2= .20, as was urban music, F(1,252) = 4.37, p = .037, ηp2 = .02, and both

gender and occupation in music industry were not, as shown in table 4. After controlling for the effects of covariates, there was no main effect found for recall. However, a small

Bonferroni post-hoc analysis shows that participants who are indeed familiar with either one or more additional verbal cues had a greater effect on song evaluation than those who were unfamiliar (Mdifference = .53, SD = .19, p = .008). However, no interaction effect was found

between recall and familiarity on song evaluation. Thus, partially accepting hypothesis 2: familiarity with additional verbal cues does cause a greater positive effect on song evaluation, however, it does not interact with recall. An overview of all found results can be found in the updated conceptual model presented in figure 2, wherein a dotted line represents no effects found.

Table 4

ANCOVA analysis: Recall of additional verbal cues and dependent variable song evaluation and familiarity as moderating variable

Main effects df F p ηp2 Mrecall MnoRecall

Recall 1, 252 .02 .879 .00

Familiarity 1, 252 7.16 .008 .03 4.17 4.14

Interaction 1, 252 .43 .512 .00

Covariates

Design evaluation 1, 252 61.07 .000 .20 Urban music preference 1, 252 4.37 .037 .02

Gender 1, 252 1.61 .206 .01

Occupation in music industry

1, 252 .40 .529 .00

Note. n = 260

Assumption A allows for further investigation in the number of additional verbal cues displayed, as in does one or more additional verbal cues cumulatively have a positive effect on song evaluation, wherein it is possible that recalling two or more cues can have a greater effect on song evaluation than for example, recalling all or one. The independent variable

‘recall of two cues’, 3 for ‘recall of three cues’ and 4 for ‘not recalling any cue’. The dependent variable was song evaluation and the variables design evaluation, urban music preference, and gender were considered as covariates. All three covariates, design evaluation F(1,253) = 61.10, p = .010, ηp2= .03, urban music preference F(1,253) = 6.72, p < .000, ηp2=

.20 and gender F(1,253) = 3.90, p = .049, ηp2= .02 were significant. After controlling for the

covariates, there was no main effect found for recall F(3,253) = 1.65, p = .179, ηp2= .02.

Thus, suggesting that there are no significant differences between the means of all four groups. However, a closer look at the means does suggest that song evaluation increases per cue as shown in table 5. Hence, while this assumption is proven to be true, the differences are not statistically significant.

Table 5

ANCOVA analysis: Recall of number of additional verbal cues and dependent variable song evaluation Main effects df F p ηp2 M SD Recall 3, 253 1.65 .179 .02 None 3.63 1.04 1 cue 3.85 1.11 2 cues 3.93 1.15 3 cues 4.27 1.32 Covariates Design evaluation 1, 253 61.10 .010 .02 Urban music preference 1, 253 6.72 .000 .20

Gender 1, 253 3.90 .049 .02

Note. n = 260.

Assumption B parallels with assumption A, as in does one or more familiar additional verbal cue cumulatively have a positive effect on song evaluation, wherein it is possible that

example, being familiar with all or one. The independent variable ‘familiarity’ is sorted between four groups wherein 1 represented ‘familiar with one cue’, 2 for ‘familiarity with two cues’, 3 for ‘familiarity with three cues’ and 4 for ‘not being familiar with any cue’. The dependent variable was song evaluation and the variables urban music preference and occupation in the music industry were considered as covariates. The covariate urban music preference F(1,254) = 4.09, p = .044, ηp2= .02 and occupation in the music industry F(1,254)

= 5.94, p = .015, ηp2= .02 were found significant. After controlling for the covariates, there is

a small significant main effect found for familiarity F(3,254) = 5.72, p < .001, ηp2= .06,

suggesting that familiarity with one or more cues significantly raises song evaluation. A Bonferroni post-hoc analysis showed that participants who are indeed familiar with two additional verbal cues had a greater effect on song evaluation than those who are unfamiliar with any cue (Mdifference = .85, SD = .24, p = .003). A closer look at the means indicates that in

fact, being familiar with two cues has a greater effect on song evaluation than none, one or all additional verbal cues as shown in table 6. Thus, this assumption that one or more familiar additional cues can raise song evaluation appears to be true, however, this effect is limited between none, one and two additional verbal cues, but not for three additional verbal cues.

Table 6

ANCOVA analysis: Familiarity with number of additional verbal cues and dependent variable song evaluation

Main effects df F p ηp2 M SD Familiarity 3, 254 5.72 .001 .06 None 3.68 1.03 1 cue 3.96 1.07 2 cues 4.52 1.44 3 cues 4.50 1.31 Covariates

Urban music preference 1, 254 4.09 .044 .02 Occupation in music

industry

The sub-question allows for an investigation whether the number of recalled and the number of familiarity with additional verbal cues have a combined and cumulative effect on song evaluation. Hence, an ANCOVA analysis was conducted with the independent variable ‘recall’ sorted between previously mentioned groups (4 groups), as was the grouped variable ‘familiarity’ (4 groups) which was added as the moderating variable. The dependent variable was song evaluation and the variables design evaluation, urban music preference, occupation in the music industry and gender were considered as covariates. Only the control variable design evaluation appeared significant F(1,242) = 57.40, p < .000, ηp2= .19, whereas urban

music preference, occupation in the music industry and gender were not, as shown in table 7. After controlling for the covariates, there was no main effect found for the number of recalled cues. However, there was a small significant main effect found for familiarity with cues F(3,242) = 2.76, p = .043, ηp2= .03. A Bonferroni post-hoc analysis showed that participants

who are indeed familiar with two additional verbal cues had a greater effect on song

evaluation than those who are unfamiliar with any cue (Mdifference = .93, SD = .33, p = .027).

However, there was no interaction effect found between recall and familiarity per additional verbal cue, as shown in table 7. A closer look at the means between each combination of recalled and familiar cue showed that the highest mean is found for two familiar additional verbal cues of which one is recalled (M = 5.00, SD = 1.73) and two familiar additional verbal cues of which none is recalled (M = 5.00, SD = 1.06), as shown in table 1 in Appendix C. Thus, indicating that two familiar additional verbal cues raises song evaluation whether one or none of the cues are recalled.

Table 7

ANCOVA analysis: Number of recalled and number of familiarity with additional verbal cues and dependent variable song evaluation

Main effects df F p ηp2 Recall (4 groups) 3, 242 .36 .785 .00 Familiarity (4 groups) 3, 242 2.76 .043 .03 Interaction 7, 242 .65 .717 .02 Covariates Design evaluation 1, 242 57.40 .000 .19

Urban music preference 1, 242 3.34 .069 .01

Gender 1, 242 1.97 .161 .01

Occupation in the music industry 1, 242 .89 .346 .00 Note. n = 260.

Additional analyses

Individual additional verbal cues. To specify which specific additional verbal cue (i.e. producer, repertoire title and artist name) is more effective in raising song evaluation and influenced by familiarity, an ANCOVA analysis was run per additional verbal cue as the independent variable sorted between those who recalled additional verbal cue and those who did not. Hence, only the relevant moderating variable familiarity was considered per analysis and song evaluation was the dependent variable in all analyses. All results are shown in table 2 in Appendix C. The results for the additional verbal cue producer showed that there are no main effects nor is there an interaction effect found on song evaluation. Thus, suggesting that there is no direct effect from recalling or being familiar with the producer on song evaluation. The results for the additional verbal cue repertoire title showed that equality for variances could not be assumed F(3,198) = 3.27, p = .022, and that there was no main effect found for recall nor an interaction effect found between recall of repertoire title and familiarity.

greater effect on song evaluation than when unfamiliar (Mdifference = .57, SD = .21, p = .009).

Lastly, the results for the additional verbal cue artist name showed that equality of variances could not be assumed F(3,93) = 3.51, p = .018, and that there was no main effect found for recall nor an interaction effect found between recall of artist name and familiarity. However, there is a small main effect found for familiarity F(1,98) = 6.46, p = .013, ηp2= .07. A

Bonferroni post-hoc analysis showed that being familiar with artist name has a greater effect on song evaluation than when unfamiliar (Mdifference = .71, SD = .28, p = .013). These results

suggest that familiarity with repertoire title and artist name have a predominantly main effect on song evaluation, whereas producer does not. In addition, none of the individual additional verbal cues interact with recall.

Figure 2. Conceptual model based on results

Discussion

The purpose of this study was to investigate the use and effect of additional verbal cues on song evaluation, and if this is influenced by familiarity with cues. The results of this study

Additional verbal cues: 1. None

2. Producer 3. Repertoire title

4. Repertoire title and artist name 5. Producer and repertoire title

6. All cues (producer, repertoire title and artist name)

Familiarity with additional verbal cues:

0 = not familiar 1 = familiar

H1+

Song evaluation

Attitude towards new song of unknown musician

H2a- sdf

reservations concerning the generalisability of the results and the extent to which the results are appropriately explained by respective literature.

First, before diving into the discussion regarding additional verbal cues, a closer look is given at the effect of visual cues. The control variable design evaluation was proven to significantly influence positive song evaluation across all analyses and explains a consistent part of the variation in song evaluation. Thus, indicating that the overall aesthetics of a music artwork cover remains effective in raising song evaluation with or without the addition of verbal cues. Furthermore, the preference for urban music was proven to effectively influence song evaluation in most analyses, which indicates that when participants are dispositioned to like urban music they are equally more dispositioned to positively evaluate the new song. These findings concur with results found in previous studies wherein a preference for similar movie genre influences the evaluation of movie success (D’Astous & Touil, 1999;

Finsterwalder et al., 2012; Sauer, 2014). Thus, suggesting that visual cues remain an integral and important part of music artwork covers in raising song evaluation. Hence, future studies should expand the current study and measure the visual elements next to verbal elements in order to accurately weigh the effects of both types of cues.

Second, the effect of additional verbal cues on music artwork covers, regardless of which specific additional verbal cue and familiarity with cues, appeared to be effective in raising song evaluation. However, this result should be interpreted with great caution as the independent variable ‘additional verbal cues’ was considered instead of the substitute

variable ‘recall’, due to the fact that ‘recall’ did not count for the condition ‘none’ which did not display any cue. This is a considerable limitation of this study given that there were no observed significant differences between all six additional verbal cue conditions and the manipulation check or recall of individual additional verbal cues. Hence, an accurate result was obscured due to the fact that the independent variable applied as an exception was not a

reliable measure of conditions with or without additional verbal cues. Nonetheless, this finding does provide a general direction for future research and is in support of previous general findings which argue that the addition of cues versus the absence of verbal cues is more effective in raising evaluation (Keller, 1987, 1991).

The third finding supports the notion that when additional verbal cues are indeed familiar, regardless of which specific cue, it can equally raise song evaluation. However, contrary to earlier studies, familiarity with additional cues did not interact with recall of additional verbal cues but appeared to have a direct effect on song evaluation. Instead, these study results coincide with earlier studies suggesting that familiarity with verbal cues can have a direct relationship on raising evaluation (Aaker & Keller, 1990; Sheinin, 2000; Swaminathan et al., 2012). A possible explanation for the lack of interaction between additional verbal cues and familiarity could be due to two reasons; first, recall of additional verbal cues was considered in the analyses as opposed to the intended ‘additional verbal cues’ variable with six conditions bearing individual or different combinations of additional cues. Thus, in fact, the interaction between recall of additional verbal cues and familiarity with cues was explored rather than the intended relationship between additional verbal cue conditions and familiarity, regardless of recall. The difference between the intended and substitute variable is that one relies on exposure to a specific condition and the other on whether participants can recall what they have been exposed to. However, exposure to a condition was not dependent on recall, whereas recall was dependent on which condition was displayed. Thus, it is possible that simply being exposed to a condition could have caused an interaction effect with familiarity regardless of recall. In fact, this is another great limitation of this study: the failed manipulation check did not allow the study to be conducted as intended and limited the analyses and results due to excluding the variable ‘additional verbal cues’ which accounts for the effect of exposure between six different conditions. Two

possible and successive explanations for the failed manipulation check is; first, the

construction of the manipulation check could have been confusing from the get-go as each answer-item grouped one or more additional verbal cues with exclusion of one or more other cues (e.g. “Reference to previously produced repertoire title by producer without mentioning artist name”). Thus, the manipulation check did not solely consist of simple answers such as ‘repertoire title’ or ‘artist name’. Second, as a result of the faulty construction, the wording of the answer-items may have been confusing to understand. Furthermore, by not having kept the answer-items simple, the actual manipulation check analysis left room for error as a few answer-items referred to one or more additional verbal cues, which may not only be

confusing in itself but also limits accurate analysis. Hence, future studies should carefully construct the manipulation check question in order to succeed and avoid having to fall back on recall of additional verbal cues. However, in spite of, these results contribute nonetheless to the body of knowledge by demonstrating that recall of verbal cues and familiarity with cues do not interact. Second, the lack of an interaction effect could be due to the fact that while familiarity may not depend on recall, recall instead can indeed influence familiarity, which may very well be the reason why familiarity had a direct effect on song evaluation and recalling of additional verbal cues did not (Kent & Allen, 1994).

Fourth, and in addition to the found direct effect of familiarity, the results of the additional analyses regarding the effect of individual additional verbal cues suggests that only familiarity with repertoire title and artist name have a direct effect on song evaluation,

whereas familiarity with producer did not. This is in stark contrast to previous studies in movie marketing wherein verbal cues referring to its counterpart (i.e. director) appears to be equally if not most effective in raising evaluation (D’Astous & Touil, 1999; Simonton, 2009). Three possible explanations for this finding could be due to; first, the fact that mentioning movie directors has been a common practice for many decades, whereas the avid mentioning

of producers on artwork covers or advertisements is fairly uncommon. Hence, this could also be the reason why familiarity of repertoire title and artist are indeed effective in raising song evaluation, simply because they are often promoted and are commonly found on artwork covers. Second, an important consideration for future research is to inspect the role of attitudes towards respective cues. It remains unclear to what extent the attitudes towards producer name, repertoire title, and respective artist influences song evaluation. As previous studies in movie marketing suggest, one of the main reasons to add cues to advertisements is the assumption that beyond familiarity, the positive attitudes towards these verbal cues raises the movie evaluation (D’Astous & Touil, 1999; Finsterwalder et al., 2012). Thus, in future studies, the role of attitudes towards respective cues should be considered to expand the current findings in this study. Third, the sample size consisted of 312 participants divided over six cover conditions which means that when sorting the participants according to those who recalled or were familiar, and those who did not, the number of participants per

combination is low. In fact, this is another great limitation of this study: the low count of participants per combination does not yield enough power to avoid type error 1 or 2 (Vanvoorhis & Morgan, 2007). Thus, while significant and non-significant effects may be found in this study, it may not hold in larger sample sizes. For example, equality of variances could not be assumed for both the direct effect analyses of repertoire title and artist name. However, in spite of, these results do provide the first clues regarding additional verbal cues on artwork covers and that in particular, the effect of repertoire title and artist name should be further explored in future studies.

Fifth, it appeared that the use of multiple cues is cumulatively more effective in raising song evaluation. This supports the notion that the use of two or more cues is indeed more effective than none or one (Laroche et al., 1996; Miyazaki et al., 2005). Furthermore, this holds true even when considering multiple familiar verbal cues. However, there appeared

to be limit of two familiar additional verbal cues although the difference between two and three cues was marginal. This marginal difference may be explained by the fact that one of the verbal cues could be the producer given that there are only significant direct familiarity effects found for repertoire title and artist name, but not for the producer. The producer, as stated previously, is not a commonly known fact. Should the producer henceforth be often promoted or gain the same levels of exposure as repertoire title and artist name, this could present different results that close the margin or even surpass it. This could indicate that like recall, more familiar additional verbal cues can increase song evaluation with each added verbal cue. Hence, future research should further explore the effects of cumulative familiar cues with a more familiar producer and repeat test the assumption. Furthermore, the number of recalled additional verbal cues and number of familiarity with cues did not interact. This could be due to similar reasons as previously explained wherein ‘recall’ was not the intended independent variable which therefore limits the analyses and results as previously described. Thus, future research could expand the testing of assumptions and sub-question by

considering ‘additional verbal cues’ as the independent variable.

In sum, the pairing of repertoire title and artist name is most commonly distributed and found, which could explain why these individual additional verbal cues were proven to have the strongest effect on song evaluation. Furthermore, this could also explain why being familiar with two cues proved more effective than all, one or none. However, when

dismissing the role of familiarity, the results indicate that three additional verbal cues are more effective than none, one or two, albeit not significantly. Thus, considering that mentioning the producer can resemble factual information in a verbal cue, which does not need to be dependent on familiarity yet still be effective in raising song evaluation, the combination of mentioning producer with a reference to repertoire title and artist name, can perhaps raise song evaluation if not even further. These findings and notion supports both

previous findings that verbal cues can consist of either factual information (Abernethy & Franke, 1996; Keller, 1987) or familiar brand names (Aaker & Keller, 1990; Janiszewski & Van Osselaer, 2000), and both types of verbal cues can contribute to raising evaluation. Assuming that referring to factual information such as who the producer is with a reference to previous work can retrieve compatible and relevant associations that can help raise song evaluation and predict benefits of listening to the new song (Keller, 1991; Meyvis & Janiszewski, 2002).

Lastly, and perhaps most interestingly, a practical application of these findings suggests that similarly to the movie industry, wherein referring to a previously directed movie (e.g. “From the director of Avatar”) proves to be equally effective in raising positive song evaluation based on familiarity with repertoire title (e.g. “From the producer of

[repertoire title]”). However, to further influence song evaluation most effectively, would be to add the respective artist name aside from repertoire title based on familiarity with the artist, and mention producer name based on factual information (e.g. “Produced by

[producer] who produced [song title] by [artist]”). Hence, a key difference between movie and music marketing with the use of additional verbal cues is that the addition of naming the producer and the artist name can further raise song evaluation. Thus, in order for the music industry to catch up with the film industry, it would require a great shift of perspective in artwork cover design. Given the current lack of additional verbal cues on music artwork covers, this suggestion for adding verbal cues could be regarded as novel or heresy for those who still regard the artwork cover as an expressive extension of their musical art. However, be that as it may, the findings from this study may have them sing to a different tune in time with regards to their next music artwork cover.

References

Aaker, D. A., & Keller, K. L. (1990). Consumer evaluations of brand extensions. Journal of Marketing, 54(1), 27–41.

Abernethy, A. M., & Franke, G. R. (1996). The information content of advertising: A meta-analysis. Journal of Advertising, 25(2), 1–17. doi: 10.2307/4188999

Bainbridge, D., Cunningham, S. J., & Downie, J. S. (2004). Visual collaging of music in a digital library. In Fifth International Conference on Music Information Retrieval (pp. 397–402). Barcelona: ISMIR.

Baker, W., Hutchinson, J. W., Moore, D., & Nedungadi, P. (1986). Brand familiarity and advertising: Effects on the evoked set and brand preference. Advances in Consumer Research, 13, 637–642.

Balachander, S., & Ghose, S. (2003). Reciprocal spillover effects: A strategic benefit of brand extensions. Journal of Marketing, 67(1), 4–13.

Baumgarth, C. (2004). Evaluations of co-brands and spill-over effects: Further empirical results. Journal of Marketing Communications, 10(2), 115–131. doi:

10.1080/13527260410001693802

Brennan, I., & Babin, L. A. (2004). Brand placement recognition. Journal of Promotion Management, 10(1–2), 185–202. doi: 10.1300/J057v10n01_13

Carpenter, G. S., Glazer, R., & Nakamoto, K. (1994). Meaningful brands from meaningless differentiation: The dependence on irrelevant attributes. Journal of Marketing Research, 31(3), 339. doi: 10.2307/3152221

D’Astous, A., & Touil, N. (1999). Consumer evaluations of movies on the basis of critics’ judgments. Psychology Marketing, 16(8), 677–694.

Dawar, N., & Anderson, P. F. (1994). The effects of order and direction on multiple brand extensions. Journal of Business Research, 30(2), 119–129. doi:

10.1016/0148-2963(94)90031-0

Desai, K. K., & Basuroy, S. (2005). Interactive influence of genre familiarity, star power, and critics’ reviews in the cultural goods industry: The case of motion pictures. Psychology and Marketing, 22(3), 203–223. doi: 10.1002/mar.20055

Elberse, A. (2007). The power of stars: Do star actors drive the success of movies? Journal of Marketing, 71(4), 102–120. doi: 10.2307/30164000

Finsterwalder, J., Kuppelwieser, V. G., & De Villiers, M. (2012). The effects of film trailers on shaping consumer expectations in the entertainment industry—A qualitative analysis. Journal of Retailing and Consumer Services, 19, 589–595. doi:

10.1016/j.jretconser.2012.07.004

Janiszewski, C., Kwee, L., & Meyvis, T. (2002). Brand alliances: The influence of brand partners on the strength of brand associations. SSRN Electronic Journal. doi: 10.2139/ssrn.1967911

Janiszewski, C., & Van Osselaer, S. M. J. J. (2000). A connectionist model of brand–quality associations. Journal of Marketing Research, 37(3), 331–350. doi:

10.1509/jmkr.37.3.331.18780

Karray, S., & Debernitz, L. (2017). The effectiveness of movie trailer advertising. International Journal of Advertising, 36(2), 368–392. doi:

10.1080/02650487.2015.1090521

Keller, K. L. (1987). Memory factors in advertising: The effect of advertising retrieval cues on brand evaluations. Journal of Consumer Research, 14(3), 316–333.

Keller, K. L. (1991). Cue compatibility and framing in advertising. Journal of Marketing Research, 28(1), 42–57.

Kent, R. J., & Allen, C. T. (1994). Competitive interference effects in consumer memory for advertising: The role of brand familiarity. Journal of Marketing, 58(3), 97. doi:

10.2307/1252313

Laroche, M., Kim, C., & Zhou, L. (1996). Brand familiarity and confidence as determinants of purchase intention: An empirical test in a multiple brand context. Journal of Business Research, 37(2), 115–120. doi: 10.1016/0148-2963(96)00056-2

Libeks, J., & Turnbull, D. (2011). You can judge an artist by an album cover: Using images for music annotation. IEEE Multimedia, 18(4), 30–37. doi: 10.1109/MMUL.2011.1 Litman, B. R., & Kohl, L. S. (1989). Predicting financial success of motion pictures: The ’80s

experience. Journal of Media Economics, 2(2), 35–50. doi: 10.1080/08997768909358184

McDaniel, M. A., & Einstein, G. O. (1993). The importance of cue familiarity and cue distinctiveness in prospective memory. Memory, 1(1), 23–41. doi:

10.1080/09658219308258223

Meyvis, T., & Janiszewski, C. (2002). Consumers’ beliefs about product benefits: The effect of obviously irrelevant product information. Journal of Consumer Research, 28(4), 618– 635. doi: 10.1086/338205

Miyazaki, A. D., Grewal, D., & Goodstein, R. C. (2005). The effect of multiple extrinsic cues on quality perceptions: A matter of consistency. Journal of Consumer Research, 32(1), 146–153. doi: 10.1086/429606

Rindfleisch, A., & Inman, J. (1998). Explaining the familiarity-liking relationship: Mere exposure, information availability, or social desirability? Marketing Letters, 9(1), 5–19. doi: 10.1023/A:1007958302123

Rodrigue, C. S., & Biswas, A. (2004). Brand alliance dependency and exclusivity: An empirical investigation. Journal of Product & Brand Management, 13(4), 477–487. Sauer, M. (2014). Cue-recognition effects in the assessment of movie trailers. Journal of

Schindler, A. (2014). A picture is worth a thousand songs: Exploring visual aspects of music. In Proceedings of the 1st International Workshop on Digital Libraries for Musicology (pp. 1–3). London: ACM. doi: 10.1145/2660168.2660185

Sheinin, D. A. (2000). The effects of experience with brand extensions on parent brand knowledge. Journal of Business Research, 49(1), 47–55. doi: 10.1016/S0148-2963(98)00116-7

Simonin, B. L., & Ruth, J. A. (1998). Is a company known by the company it keeps?

Assessing the spillover effects of brand alliances on consumer brand attitudes. Journal of Marketing Research, 3553(15), 30–4243.

Simonton, D. K. (2009). Cinematic success criteria and their predictors: The art and business of the film industry. Psychology & Marketing, 26(5), 400–420. doi: 10.1002/mar

Swaminathan, V., Reddy, S. K., & Loughran Dommer, S. (2012). Spillover effects of

ingredient branded strategies on brand choice: A field study. Marketing Letters, 23, 237– 251. doi: 10.1007/s11002-011-9150-5

Van Osselaer, S. M. J., & Janiszewski, C. (2001). Two ways of learning brand associations. Journal of Consumer Research, 28(2), 202–223. doi: 10.1086/322898

Vanvoorhis, C. R. W., & Morgan, B. L. (2007). Understanding power and rules of thumb for determining sample sizes. Tutorials in Quantitative Methods for Psychology, 3(2), 43– 50. doi: 10.20982/tqmp.03.2.p043

Washburn, J. H. (2000). Co-branding: brand equity and trial effects. Journal of Consumer Marketing, 17(7), 591–604.

Appendix A

Cover conditions with respective cues

1. None

2. Producer

3. Repertoire title

4. Repertoire title and artist name

5. Producer and repertoire title

5. Producer, repertoire title and artist name

Appendix B

Manipulation check question

Which verbal elements did you see on the music artwork cover? Multiple answers are possible, please tick what you think you’ve seen based on previously shown music artwork cover

▢

▢

▢

▢

▢

▢

Appendix C Tables Table 1

Means and SD between number of recalled and familiar additional verbal cues groups

Familiarity Recall Mean SD n

1 cue familiar 1 cue recalled 3.95 1.10 41

2 cues recalled 4.15 1.17 12

3 cues recalled 3.50 .66 3

No cues recalled 3.89 .92 7

2 cues familiar 1 cue recalled 5.00 1.73 3

2 cues recalled 4.35 1.37 18

3 cues recalled 4.88 3.00 2

No cues recalled 5.00 1.06 2

3 cues familiar 2 cues recalled 2.75 . 1

3 cues recalled 4.66 1.25 11

No cues familiar 1 cue recalled 3.77 1.08 93

2 cues recalled 3.62 .91 29

3 cues recalled 3.65 .91 5

No cues recalled 3.49 1.02 33