Nudging with Music. The Influence of Associated Music, in combination with Volume, on Product Choice in a Virtual Supermarket.

Transcription

1 Nudging with Music The Influence f Assciated Music, in cmbinatin with Vlume, n Prduct Chice in a Virtual Supermarket Wageningen University Master Management, Ecnmics and Cnsumer studies Chair grup: Marketing and Cnsumer Behaviur (MCB) Fd and Bibased Research (FBR) Curse cde: MCB Date: March 2015 Authr: Jasper Freriks Cntact: jasper.freriks@wur.nl Student number: th supervisr: dr. H.W.I. (Erica) van Herpen 2nd supervisr: A.J. (Anna) Maaskant Keywrds: nudging, assciated music, vlume, virtual supermarket, prduct chice, prduct recgnitin.

2 Preface In frnt f yu lies the result f 6 mnths f wrk, my MSc thesis Nudging with Music: The Influence f Assciated Music, in cmbinatin with Vlume, n Prduct Chice in a Virtual Supermarket. This MSc Thesis is part f the master prgram Management, Ecnmics & Cnsumer Behaviur at Wageningen University. I wrte this thesis at the chair grups Marketing and Cnsumer Behaviur (MCB) and the research institute Fd and Bibased Research (FBR). My warm thanks g t my supervisrs Erica van Herpen (MCB) and Anna Maaskant (FBR) fr their feedback as well as supprt during my thesis perid. In specific, during the analyses their help was mre than useful. Further I wuld like t thank Thmas van Straaten (Md Media) fr his advice cncerning the music sngs fr the virtual supermarket. Als many thanks t Daniella Stijnen (FBR) and Ilse Plet (FBR) fr their help with the mvement f the virtual supermarket. Furthermre, I wuld like t thank Arnud Rrda and Niclien van der Waa fr their help with the recruitment f participants. Last but definitely nt least, I wuld like t thank all participants f my experiments. Jasper Freriks Wageningen, March

3 Summary In the cnsumer decisin-making prcess, cnsumers respnd t mre than just the prduct itself (Billings, 1990). Als stimuli in a certain envirnment can have an effect n the feelings and behaviur f cnsumers (Mehrabian, 1976). This effect f stimuli can take place in several ways, fr example by nudges. Nudges are subtle ways f steering cnsumers t the mst desirable utcme (Lewis, 2008). In the DONRO-prject music is ne f the nudging strategies that will be researched. The effect f music, that is assciated with a prduct categry, is aimed t be identified. Als the effect f vlume f this music is aimed t be identified. Literature abut music in a retail envirnment is limited. Within the existing literature tw imprtant theries cncerning the effect f music n cnsumer behaviur exist. The first thery is the stimulus rganism respnse mdel by Mehrabian and Russel (1974) and fcusses n emtinal states (pleasure, arusal, and dminance). The secnd thery is the assciatin thery and fcusses n intercnnected cgnitive cncepts. This assciatin thery als fcusses mre, cmpared t the Mehrabian and Russell mdel (1974), n the final purchase. In literature is fund that sngs that were strngly assciated with a cuntry had an effect n prduct purchase. These sngs resulted in buying relatively mre wine frm the assciated cuntry (Nrth et al., 1999). Hwever, besides this study there is little research dne yet n the influence f a strngly assciated sng n prduct chice in a retail envirnment. Next t that, the effect f vlume f this music is unknwn. This research aims t gain mre insights n the effect f musical nudges. These nudges shuld ultimately lead t healthier and mre sustainable fd chice, which is relevant fr the sciety. Frm the pre-tests was cncluded that the Asian sng (Dansngs Ambershire), Mediterranean sng (Italian Medley) and neutral sng (Easy Listening Music: Guitar & Sax) were mst useful. The Asian and Mediterranean sng have the strngest assciatins with respectively Asian and Mediterranean meals versus ther meals. Therefre, these sngs have been used as nudge in the virtual supermarket and have been cmpared with neutral music. In the virtual supermarket 10 Asian, 10 Mediterranean and 10 ther meal packages were used. 294 participants were divided in ne f the cnditins frm a 3 (sng type: Asian, Mediterranean, neutral) x 3 (vlume: lud, medium, sft) between subjects design. The hyptheses shwed that it was expected that a strngly assciated sng leads t mre sales f the assciated meal packages. Next t that it was expected that thse assciated meal packages wuld be better recgnized in case the assciated sng was played. N significant results f the sngs n prduct chice and prduct recgnitin (memry) were fund. Besides, vlume f music did nt have an influence n bth prcesses (prduct chice and prduct recgnitin). In cnclusin, this research seems nt t supprt the thery that strngly assciated music can influence prduct chice and recgnitin (memry). This research can be cnsidered as a useful starting pint fr future research cncerning the effect f music n assciated prduct categries. As described in the discussin, it might be that n significant results were fund because the music was cnsidered as t inapprpriate. Future research are advised t make use f sngs that are mre apprpriate in a supermarket envirnment. Next t that, it might be that cncerning meal packages participants had a strng pinin. It culd be that ther prducts r ther prduct categries are easier t nudge. This culd be taken int accunt when ding future research. 2

4 Table f Cntents 1. Intrductin Backgrund Relevance and aim Research questins Literature backgrund Assciatins... 8 Effect f typical sngs n cnsumer behaviur... 8 Assciatin and knwledge structure thery Effect f vlume f music n prduct chice Wrking memry and prduct recgnitin Methdlgy Results f a previus pre-test and pilt study Pre-test sngs Pre-test sngs Pre-test vlume Try-ut Virtual supermarket Results virtual supermarket Statistical analysis Sample Similarity cnstrains Manipulatin check Results Cnclusin Discussin Limitatins and future research Implicatins References Appendixes

5 1. Intrductin Nwadays it is knwn that, within the cmplex cnsumer decisin-making prcess, cnsumers respnd t mre than just the prduct itself r service being ffered (Billings, 1990). Studies f psychlgists shw imprtant envirnment-behaviur relatinships, resulting in a grwing psychlgical discipline called envirnmental psychlgy (Dnvan & Rssiter, 1982). This discipline fcusses n the effect f stimuli in a certain envirnment n the feelings and behaviur f cnsumers (Mehrabian, 1976). Stimuli are aspects that have an impact r effect n a persn s that the behaviur is mdified r changed (Bitner, 1992). The effect f stimuli in a certain envirnment n the feelings and behaviur can take place in several ways, fr example by nudges. Lewis (2008) described nudges as subtle ways f steering cnsumers t the mst desirable utcme. In retail envirnments, like a supermarket, cnsumers can be influenced by many different types f stimuli. Amng many in-stre elements, music is ne f the stimuli factrs that can exert influence n cnsumers. Many articles shwing effects f music n certain emtins, beliefs and physilgical sensatins, already exist (Bitner, 1992; McGldrick & Piers, 1998; Hsea, 2004; Alpert & Alpert, 1990). Anther way in which music can be used (as a nudge) is by assciatins f that music. When music is heard, assciatins can cme up in the mind f a cnsumer. These assciatins culd influence cnsumer behaviur and is therefre f relevance. This because this steering methd (nudge) culd be applied fr prducts r prduct categries that are healthy and/r sustainable, this in rder t stimulate purchases f these prducts r prduct categries. Nwadays, health and sustainability are imprtant tpics, because f the increasing prblem f besity (Wyatt et al., 2006) and the need t behave in a sustainable way t ensure a liveable envirnment fr next generatins (Degenhardt, 2002). This study is part f the DONRO-prject, which means: Develper f Nudges fr Retail and Out-f- Hme. The DONRO-prject has as main gal t measure the direct effect f nudges n individual cnsumers. This in rder t measure pssible changes in chice behaviur. Retail as well as catering cmpanies are invlved in this prject. This research will specifically fcus n nudging with music in a retail envirnment. 1.1 Backgrund Tw imprtant theries cncerning the effect f music n cnsumer behaviur exist and are discussed in this sectin. The first thery that is discussed belw, is the stimulus rganism respnse mdel by Mehrabian and Russel (1974) and fcusses n emtinal states (pleasure, arusal, and dminance). The secnd thery is the assciatin thery and fcusses n intercnnected cgnitive cncepts. This assciatin thery als fcusses mre, cmpared t the Mehrabian and Russell mdel (1974), n the final purchase. One way in which music can influence cnsumer behaviur is via primary emtinal respnses. This can be explained by the classical envirnment mdel by Mehrabian and Russell (1974). The Mehrabian- Russell apprach uses a stimulus rganism respnse mdel. Music culd act as a stimulus (frm the envirnment) in a stre. Thus, music culd influence the rganism and, in specific, the emtinal state. Mehrabian and Russell's mdel states that three basic emtinal states (pleasure, arusal, and dminance) create all the feelings a persn culd have. This framewrk 4

6 suggests that a certain envirnmental stimulus, like music, culd cause a particular individual emtinal respnse, which, in turn, influences the individual. As a behaviural respnse, a persn culd apprach r avid the envirnment t a relatively greater r lesser degree (Mehrabian, 1976). Accrding t Mehrabian and Russell (1974) appraching behaviur cnsist f physically mving tward, explring, cmmunicating, and perfrming in an envirnment. Avidance behaviur leads t a desire t leave, disinterest, lack f interactin, and pr perfrmance in an envirnment. A different thery, which is discussed mre recently, cncerning the influence f music n cnsumer behaviur, is thrugh knwledge activatin and cnnectins. Martindale and Mre (1988) argued that a persns mind cnsists f intercnnected cgnitive cncepts. Thse cncepts are cnnected with each ther by assciatin. Whenever a persn thinks abut a cncept, the persn will als think abut related aspects. Cncepts differ in strength in which they can be activated. Cncepts that are assciated frequently are strnger than assciatins that are nt made ften. Which cncepts becme activated depends als n the availability and accessibility f thse cncepts (Niedenthal et al., 2011). In case a cncept is recently activated, it is pssible that this influences the interpretatin f later encuntered infrmatin. Within the psychlgical science, this prcess is called priming (Kunda, 1999). In case f music, certain music can pssibly activate related knwledge structures. When music in stres leads t an activatin f relevant related knwledge structures, this might als influence purchase decisins. Literature supprts this thery, since Areni & Kim (1993) shwed that cnsumers bught mre expensive wine in a wine stre when classical music was the backgrund music, cmpared t pp music. This is prbably because peple assciate the (classical) music with mre expensive prducts (Oakes, 2000). Nrth et al. (1999) fund that sngs that were strngly assciated with a cuntry had an effect n prduct purchase. These sngs resulted in buying relatively mre wine frm the assciated cuntry. Tw imprtant theries cncerning the effect f music n cnsumer behaviur are discussed. Althugh bth theries can be used t study the effect f music n cnsumer behaviur, the theries differ greatly with regard t the explanatin. As shwn, the Mehrabian- Russell apprach uses a stimulus rganism respnse mdel, while the secnd thery fcusses n assciatins and the knwledge structure. In the next sectins, this thesis will fcus n the assciatin and knwledge structure thery, because this (mre recent) thery is able t prvide new insights cncerning the effect f music n cnsumer behaviur. The new insights are explained in the sectin Relevance and aim Since mst studies studying music as a stimulus make use f the mdel by Mehrabian & Russell (1974), the knwledge structure and assciatin thery is nt fully studied. This is because the Mehrabian & Russell (1974) mdel des nt explicitly include assciatins with a certain prduct r prduct categry in their mdel. Mrever, previus studies that d use the assciatin thery, nly apply the thery at the wine sectin categry (Areni & Kim, 1993; Nrth et al., 1999). S, n research is dne, cncerning the effect f music and the activatin f an assciatin in ther prduct grups than wine. As explained, certain music culd influence whether r t what extent an activatin f assciatin ccurs. Besides this, within the assciatin thery, the effect f vlume f music is nt studied. Previus research shws that vlume can have majr effects (Garlin and Owen, 2006; Smith & Curnw, 1966). Fr example, a sft vlume led t a lnger stay cmpared t a luder vlume (Smith & Curnw, 1966). Althugh these studies are dne in a stimulus rganism respnse mdel, 5

7 in this study it is expected that als in the assciatin and knwledge thery vlume can play an imprtant rle. It is pssible that a higher r lwer vlume increases the activatin r strength f an assciatin with a certain prduct r prduct categry. This increase f activatin r strength f an assciatin, culd wrk in varius ways. It culd be the case that a higher vlume results in mre awareness f the music. Which, in turn, culd lead t a greater chance r strnger effect f assciating related prducts r prduct categry. The increase f assciatins with related prducts, culd lead t mre sales f thse prducts. Hwever, music culd als lead t an unpleasant feeling, resulting in a desire t leave. Anther pssibility is that peple becme mre aware f the nudge which culd result in resistance. Existing studies mainly fcus n the utcmes f affective variables f cnsumer behaviur (Dubé & Mrin, 2001; Alpert et al., 2005; Alpert & Alpert, 1990), while in this thesis the fcus n cnsumer behaviur cnsists f tw aspects. First, the effect f music n the sales f the test prducts (prduct chice); secnd, the effect f music n prduct recgnitin (prduct memry). The effect n prduct sales is f imprtance since results f this aspect culd shw whether an assciatin f music with certain prducts r prduct categry (due t music) als leads t actual purchases f these prducts. Prduct recgnitin (memry) is als relevant since it culd be that music that is assciated with a certain prduct r prduct categry des nt lead t an increase f the sales f thse prducts, but des create mre attentin t these prducts. This attentin can be measured by prduct recgnitin, since literature has shwn that recgnitin can be used as an indicatr f attentin (Pieters et al., 2002). Mre attentin will lead t a better recgnitin f prducts. This increase f attentin (and thus recgnitin) culd be a beginning f, ultimately, mre (assciated) prduct r prduct categry purchases, as is explained in chapter 2.3. Next t this knwledge gap, mre infrmatin abut the ptential influence f musical assciatin by music, and the musical characteristic vlume, n cnsumer behaviur is als f practical relevance. When it is shwn that an assciatin f music can lead t different prduct chices, these nudges can be applied t stimulate purchases f ther prduct categries (fr example healthy and/r sustainable fd). Thus, since the effect f vlume cmbined with music that is strngly assciated with a certain prduct r prduct categry is unknwn, a new useful research culd be carried ut. A cnceptual mdel f the research is shwn in figure 1. This figure shws that music can lead t an assciatin f a prduct r prduct categry. The activatin r strength f an assciatin n a prduct r prduct categry culd differ per sng, therefre tw different sngs will be used. Vlume can als play a rle in this assciatin. Music vlume is expected t have an influence in the activatin prcess. As explained, it is pssible that a higher (r lwer) vlume increases the activatin r strength f an assciatin with a certain prduct r prduct categry. Hwever, it shuld be taken int accunt that the effect f vlume n the assciatin culd differ per sng, since a certain level f vlume culd fit better in a certain sng cmpared t anther sng. The derived assciatin can influence prduct recgnitin (memry) and prduct chice. 6

8 Figure 1: Cnceptual mdel f research. 1.3 Research questins Main questin: What is the influence f assciated music, and the musical characteristic vlume, n prduct chice and prduct recgnitin (memry)? Sub questins: What is the influence f music, which is assciated with a certain prduct r prduct categry, n prduct chice and prduct recgnitin? What is the influence f vlume n prduct chice and prduct recgnitin? In the next chapter literature is discussed. By making use f this literature, hyptheses are derived. Chapter 3, the methdlgy, describes hw the hyptheses are tested. In chapter 4 the results f this research is discussed. Finally, cnclusins and the discussin are presented in chapter 5. 7

9 2. Literature backgrund First, the effect f typical sngs n cnsumer behaviur and the assciatin thery is described. Thereafter, literature cncerning vlume is discussed. Althugh, these studies make use f the classical envirnment mdel (Mehrabian & Russel, 1974), the main results and findings can be cnsidered as relevant backgrund infrmatin. At last, wrking memry and prduct recgnitin literature is discussed. In this literature backgrund, literature cncerning cnsumptin and purchases are used. T make clear what the differences are between cnsumptin and purchases, the fllwing distinctin is used: cnsumptin is the act r prcess f cnsuming by eating r drinking, while a purchase is when a gd r service is bught. Althugh this thesis will fcus n cnsumer behaviur in a retail envirnment (virtual supermarket), theries and literature discussed abut cnsumptin will als be used. These are used because literature cncerning purchase behaviur is limited. Literature using cnsumptin is als relevant since the theries cncerning fd intake als deal with influences frm the atmsphere, like music. 2.1 Assciatins In this part, literature that supprts that music can be assciated with certain prducts r prduct categries, is explained. Thereafter is explained hw these musical assciatins wrk in a cnsumers mind. Effect f typical sngs n cnsumer behaviur Different typical sngs can lead t different prduct chices (Agmn, 1990). Certain types f backgrund music may influence cnsumers t spent mre expensive prducts, as shwn by Areni & Kim (1993). They discussed that cnsumers bught mre expensive wine (in a wine stre) when classical music was heard, cmpared t a cnditin in which pp music was heard. This is a result f the linkage and assciatin f classical music with expensive prducts (Oakes, 2000). A study that supprts this thery, is dne by Nrth et al. (1999). They fund that sngs that were strngly assciated with a cuntry had an effect n prduct chices. These sngs resulted in buying relatively mre wine frm the assciated cuntry. In specific, French music resulted in relatively mre purchases f French wine and German music f German wine. Hwever, Wilsn (2003) shwed that classical music des nt necessarily leads t mre cnsumptin r purchasing. In the research was fund that peple cnsumed mre beverages in a restaurant when jazz, easy listening r ppular music was played, cmpared t classical music. This because classical music did nt seem t fit in the cntext. These findings supprt the thery frm Radcy and Byle (1997), they suggested that peple spend mre mney and time in a stre r restaurant in case the music played was cnsidered as apprpriate. Similarly, MacInnis and Park (1991), discussed that peple are easier t persuade in case the music is apprpriate fr the cntext. Thus, musical assciatins can influence the prduct chice f cnsumers. A prerequisite is that the music fits within the cntext. Hw these musical assciatins wrk in a cnsumers mind, is explained in the next sectin. 8

10 Regarding the effect f different typical sngs n prduct recgnitin, a gap in literature exists. Hwever, literature cncerning prduct recgnitin and wrking memry des exist and is f imprtance. Therefre, a sidestep is made and literature cncerning wrking memry and prduct recgnitin is discussed in sectin 2.3. Assciatin and knwledge structure thery Since this study makes use f the assciatin and knwledge structure thery, this thery is discussed. In the intrductin, this thery is already shrtly intrduced, but in this part it is explained in detail. A lt f infrmatin is perceived in daily life, peple classify this infrmatin in cncepts. A cncept is a mental representatin f a categry, this categry is an amunt f bjects (r attributes) that we think that belng tgether (Smith, 1990). Thse cncepts can cntain abstract as well as cncrete bjects. Fr example, when ne things abut Italy, pizza and pasta may be assciated, but als happiness because f yur last hliday in Rme. The cncepts allw us t classify things (like behaviur, events and peple), t interpret ur wrld and make implicatins that are nt directly bserved. All cncepts tgether, frm a framewrk, which help us in daily life. This because it helps us t classify, with cmmunicating, reasning, guiding attentin and interpretatin (Bargh, 2006; Kunda, 1999). Within the assciatin and knwledge structure thery, different views exist, namely: classical, prbabilistic and cncepts as thery based view. In the classical view, the main assumptin is that cncepts cnsist f necessary and sufficient attributes. Defining fr example the cncept banker ne may be cnsidered as a banker if he r she has all f the attributes. Mrever, if anyne has all the attributes, ne must be a banker. After sme critique cncerning this view, the prbabilistic view came up. Wittgenstein (1953) stated that a categry can be described by a list f attributes that are typical f it, but nt necessarily define it. As a clear example he used birds; they fly and nest in trees, but a chicken is als a bird and des nt shw this behaviur. The thery-based view is mstly similar t the prbabilistic view, but it adds that many cncepts and attributes are cnnected because f causal relatinships (Keil, 1989). Fr example, when ne thinks abut besity, ne culd prbably think f unhealthy fd. This because unhealthy fd can lead t besity; there is a causal relatinship. Since the thery based view is the mst cmprehensive, this view is used. Amng the many cncepts, sme may becme salient and thers nt. This depends n tw aspects, namely accessibility and availability. Availability is abut the framewrk that exists in the mind f a persn. Whether thse cncepts are available depends n earlier experiences and the cnnectins that are made in the past between aspects that belng t the cncept. Because experiences play a large rle in this assciatin thery, the cncepts that are frmed are persnal and are applicable fr a persn nly. Hwever, sme aspects are experienced by a lt f peple, fr example a lt f peple will assciate Italy and pizza (Niedenthal, 2011). 9

11 Accessibility is the pssibility t activate available knwledge. Whether a cncept becmes activated depends n three factrs, namely: the stimulus, the cntext and the persn wh perceives it. The activatin f the cncepts can be influenced by a stimulus. Fr example, when ne sees a bald head f a men ne may interpret this persn as a bit lder persn wh lst his hair, but when this is a wman with a bald head it may be interpreted with a certain illness. Thus, this interpretatin f a stimulus can activate different cncepts. Cntext can als have an influence n the prcess f activatin. This phenmena is called cntextual priming (Mrris & Flk, 1998). An example f cntextual priming is shwn by a study f Berger et al. (2008). They discussed that the place where peple vte had an influence n their eventual vte. The persn wh perceives it, r called the bserver, als influences cncepts that are activated. This because their persnal framewrk is frmed by their wn life (Kunda, 1999). Fr example, smene wh likes the Nkia phnes, will recgnize thse phnes in less time and needs les input than smene wh des nt like r is nt familiar with Nkia phnes. Next t that, peple are als mre likely t activate cncepts that help t achieve their gals (Kunda, 1999). There are three types f knwledge representatin mdels that shw hw cncepts are represented and interrelated. These mdels als describe hw the knwledge activatin and retrieving prcess wrk. These are the cnsensus, assciative netwrk and parallel-cnstraint-satisfactin mdel. A cnsensus map is a tl t lk int the mind f a cnsumer and shws the uncnscius prcess. Thse maps shw shared thughts and/r feelings by cnsumers abut prducts, next t that they als represent general cncepts. Als the interactin between the cncepts f a prduct can be shwn in a cnsensus map (Zaltman, 2003). Figure 2: Example f a cnsensus map (Zaltman, 2003). Next t thse cnsensus maps, als assciative netwrk maps exist. The cnsensus maps fcus n thughts and feelings abut a prduct, but this assciatin map fcusses n the assciatin with a certain prduct r prduct categry. These assciative netwrk maps (see figure 3) shw mental representatins as netwrks f ndes. Ndes are features, cncepts r prpsitin and thse ndes are linked by assciatins. When a certain nde is perceived, this nde can be activated and als related ndes can becme activated. S, if ne thinks abut a certain nde, the chance f activatin f related ndes increases. Every nde has its wn level f activatin, since it depends n the ccasin. In case a certain nde is mre activated, it is mre likely peple becme aware f that 10

12 nde. Next t that, when a nde is mre times activated, it is als mre likely that the nde will be recalled and that it will be applied t incming infrmatin. In case f a strng assciatin, related ndes will be activated easier and faster (Kunda, 1999). In case a nde is activated recent, it can influence interpretatin f infrmatin which is perceived later n. This prcess is als called priming (Bargh, 2006). After being activated, the level f activatin f a nde decays ver time (Kunda, 1999). Figure 3: Example f an assciative netwrk, case: McDnalds (French & Smith, 2013). The Parallel-cnstraint-satisfactin mdel has a lt in cmmn with the assciative netwrk mdel, but add that when ndes are activated, ndes can nt nly active related ndes, but als deactivate them (Rumelhart & McClelland, 1986). In case ndes have a psitive link, activatin f a nde will increase activatin f the ther nde. But, in case ndes have a negative link, activatin f ne nde will lead t a decrease f activatin f the ther nde. Using figure 3, a link between Children s parties and Cheap can be added, but a link between Children s parties and unhealthy is less likely t be activated. This because, mst f the times, cheap is seen as smething psitive, while unhealthy is seen as negative. The parallel-cnstrain-satisfactin mdel can be divided in tw mdels that have a different level f abstractin f the cnceptualized ndes. Thse classes are lcal representatins and distributed representatins (Kunda, 1999). Lcal representatins are similar t assciative netwrk maps, as described n the previus page, with additin f the (de-)activatin. Distributed is mre hierarchical and the ndes are in this mdel mre basic, lwer level elements. Thus, a cncept cnsists f a number f ndes t becme clear. The distributed representatin is fr example used in rder t recgnize a wrd and is less relevant in this study. In this thesis, the assciative netwrk mdel is mst applicable. This because the distributed representatin f the parallel-cnstraint-satisfactin mdel is nt useful, since this is t hierarchical, this is mre useful fr wrd recgnitin. Next t that the assciative netwrk mdel fcusses n assciatins and nt nly n thughts and feelings, like a cnsensus map. Mrever, the assciatin thery fcusses n effects f assciatin, which is f relevance. In this thesis an assciatin effect culd be that music culd influences prduct chices. 11

13 Making this assciatin effect mre specific, it is likely that music that is assciated with a certain prduct r prduct categry leads t mre purchases f this prduct r prduct categry. A cmparisn can be made between music that is assciated with a certain prduct r prduct categry with music that is nt assciated with a certain prduct r prduct categry. Taken abve infrmatin int accunt, hypthesis 1 is derived. Hypthesis 1: When cnsumers listen t music that is assciated with a prduct r prduct categry, this will lead t mre prduct purchases f that prduct r prduct categry cmpared t when cnsumers listen t music that is nt assciated. 2.2 Effect f vlume f music n prduct chice In this part the effect f vlume f music n prduct chice is discussed. Prduct chice is the sales f the test prducts. Studies cncerning the effect f the musical characteristic vlume shw different results. It shuld be nted that thse studies mainly fcus n certain emtins, beliefs and physilgical sensatins. The studies d nt make use f the assciatin thery, as explained in 1.1. Within the aspect prduct chice studies have shwn that in case vlume is relatively high, cnsumers will spend less time in a supermarket. Hwever, the rate f spending per minute in a stre will increase cmpared t a lwer vlume (Smith & Curnw, 1966). Garlin and Owen (2006) suggested that a lwer music vlume leads t a lnger stay and, ultimately, mre purchases. Herringtn and Capella (1996) fund n effect f vlume n shpping expenses. Cncerning cnsumptin, literature shws that fd cnsumptin seems t increase when lud and fast music is played (McCarrn & Tierney, 1989), while n the ther hand slwer and sfter music leads t a lnger stay which als leads t mre cnsumptin (Strebele & De Castr, 2004). Sullivan (2002) mentined that sft music leads t mre fd intake cmpared t lud music. Hwever, Sullivan (2002) als nted that there might be a medium amunt f db in which intake culd be ptimized. T make the vlume sectin mre clear, table 1 is prvided n the next page. Within this table the effect f an increase f vlume n prduct chice / fd intake, is shwn. The first 3 studies (Garlin & Owen, 2006; Herringtn & Capella, 1996; Smith & Curnw, 1966) are cncerning purchases and the last 3 studies are cncerning cnsumptin (Sullivan, 2002; Strebele & De Castr, 2004; McCarrn & Tierney, 1989). 12

14 Dependent Variable Reference Genre Envirnment Vlume Prduct Chice r Fd Intake Garlin & Owen, Varius (n specific Retail Sft / Lud (N db PC: NS 2006 Herringtn & Capella, 1996 Smith & Curnw, 1966 Sullivan, 2002 Strebele & De Castr, 2004 McCarrn & Tierney, 1989 genre mentined) Varius (n specific genre mentined) Backgrund (N specific genre mentined) Backgrund (N specific genre mentined) Backgrund (N specific genre mentined) Backgrund (N specific genre mentined) Supermarket Supermarket Restaurant Restaurant Restaurant mentined) Sft / Lud (N db mentined, supermarket manager decided) Sft / Lud (N db mentined, supermarket manager decided) 95 per cent f the sample agreed that the vlume was either sft r lud Sft / Lud (N db mentined) Sft: 70 db Lud: 90 db Table 1: The effect f the independent variable vlume f music n prduct chice and stay duratin. PC: N effect PC: N effect FI: - FI: NS T summarize, the effect f a higher vlume n prduct chice is unclear. This because ut f the six studies, fur different results are fund. This culd be the case because different studies used different amunts f decibels r did nt even mentin the amunt but just mentined luder music. The thught that the amunt f decibels culd play a rle can be cmbined with previus research dne by Vitz (1972). He shwed that a preference fr tne intensity (db) has the frm f an inverted U. In this inverted U, sft music is preferred abve lud, but a in between ptin leads t ptimizatin f preference Cmbining the assciatin thery and literature regarding vlume, it can be stated that vlume als has an influence n prduct purchases. As shwn by Sullivan (2002) a lw vlume is preferred abve a high vlume. It is assumed that a (t) high vlume culd lead t an unpleasant feeling, because ne culd cnsider the music as t lud, resulting in a desire t leave. Next t that, because f the reduced pleasantness, a persn s mind culd be less pen fr assciatins. Anther pssibility, because lud music stands ut, is that peple becme mre aware f the nudge which culd result in resistance. Because f this resistance, peple culd decide t nt buy anything that is assciated with the music. This in rder t make a pint against being cntrlled. Sullivan (2002) als nted that there might be a medium amunt f db in which intake culd be ptimised. This thught can be cmbined with the inverted U f Vitz (1972). It is assumed that when peple feel pleasant (like in the tp f the inverted U), peple are mre pen fr assciatins. It culd be the case that a middle vlume results in mre awareness f the music. Which, in turn, culd lead t a greater chance r strnger effect f assciating related prducts r a prduct categry, and thus purchases f these prducts. Thus, prduct sales f the assciated prducts r prduct categries will als increase in this middle phase. S, it is expected that the ludness f music results in mre awareness, which has a psitive influence n the assciatin effect. This prcess wrk until a certain pint, in which FI: + 13

15 music becmes t lud and this has several negative influences. Making use f this literature, hypthesis 2 is derived. Hypthesis 2: The effect f vlume n purchase f prducts (assciated with the music) will have the frm f an inverted U 2.3 Wrking memry and prduct recgnitin The effect f music that is assciated with certain prducts, culd directly lead t purchases f thse prducts as well as in an indirect way. In this indirect manner, the viewing behaviur is f imprtance and this is explained in this sectin. This attentin-getting prcess is best described by Duncan and Humphreys (1989). They described that visual infrmatin is stred r maintained in wrking memry. A clse interactin between wrking memry and attentin exist. Dwning (2000) shwed nt nly that what has been seen (attentin) is stred in wrking memry, but als that the infrmatin prcessed in wrking memry influences what is attended in the visual scene later n. Aspects that are already prcessed nce, will stand ut later n. Thus, a prduct which is prcessed in wrking memry (due t attentin) will stand ut later n. This standing ut, culd result in mre attentin which, in turn, might result in a purchase. S an increase in attentin, culd lead t a direct purchase, but culd als lead t a purchase later n. Hwever, it shuld be nted that the duratin f this standing ut effect is limited. This limitatin differs per prduct and persn (Dwning, 2000). In case a certain lcatin r bject gets attentin, prcessing f it is enhanced, while at the same time prcessing f nt selected lcatins and/r bjects is suppressed (Pieters & Wedel, 2004). Tw mechanisms play a rle in this prcess, namely bttm up and tp dwn mechanisms. Bth mechanisms have an influence n the visual attentin. Bttm up evlves frm the visual envirnment, fr example by visual stimuli. Shrtly, it is emanating frm the stimuli that a cnsumer is expsed t (Pieters & Wedel, 2006); attentin can be captured by features f a stimulus that are salient, fr example by clur, luminance r rientatin, r the appearance f new bjects (Jnides & Yantis, 1988; Yantis & Hillstrm, 1994). In particular, irregularities frm the glbal image structure lead t a bttm up mechanisms. The manipulatin f features means this type f search can be efficient and prviding quick reactin times by creating a pp ut effect (Treisman & Gelade, 1980). This pp-ut effect can nly ccur in the bttm-up mechanism, in this effect an dd item gets attentin immediately, independent f the present distractrs (Pieter & Wedel, 2006). On the ther hand, tp dwn prcesses can steer attentinal guidance. This prcess is evlving frm the behaviural aspects f visual scenes and riginates fr example frm gals, memry and states and traits f a persn. This culd be in case cnsumers pay attentin t specific spts r regins (Psner, 1980), r when they direct attentin twards certain individual bjects (Duncan, 1984; Rck & Gutman, 1981) r features f bjects (Bacn & Egeth, 1994). There is n pp ut effect like in a bttm up prcess, therefre it is harder t identify the prducts fr cnsumers. Mstly, this prcess takes mre time and can be cnsidered as less efficient cmpared t the bttm up prcess (Pieters & Wedel, 2006). 14

16 T measure viewing behaviur, the extent t which participants culd recgnize a prduct afterwards can be used. Literature shws that recgnitin f prducts is strngly influenced by the extent t which ne maintains a visual fcus n a single place (eye fixatin) (Lftus, 1972; Lftus et al., 1991; Pieters et al., 2002). Mrever, Danaher & Mullarkey (2003) fund that web page expsure duratin is a gd indicatr f banner advertising recgnitin afterwards. Furthermre, Rethans et al. (1986) shwed that recgnitin f advertising cntent increased when the frequency f expsure increased. The assciatin effect that is explained in 2.1, will als have influences n the viewing behaviur. The music played, culd result in an assciatin with a prduct r prduct categry. Music culd create a tp dwn effect, as just described. In this effect attentinal guidance is dminated and it culd be that cnsumer create r have spts r regins which they attend t (due t music). Fr example, it culd be that Italian music leads t attentin fr pizza ut f the freezer. It can be hypthesized that these prducts will therefre get mre attentin cmpared t situatins with unassciated music. This increase in attentin, culd lead t a better prduct recgnitin as shwn by Pieters et al. (2002). Belw, hypthesis 3 is derived. Hypthesis 3: When cnsumers listen t music that is assciated with a prduct r prduct categry, these prducts r prduct categry will be better recgnized cmpared t when cnsumers listen t music that is nt assciated. As described in hypthesis 3, attentin and thus recgnitin will increase with the assciated prducts. Hwever, Pieters and Wedel (2004) als shwed that in case a certain bject is selected by attentin, prcessing and paying attentin f it is enhanced, while at the same time prcessing and attentin f nt selected lcatins and/r bjects is suppressed. S, an increase f attentin twards assciated prducts, will lead t a decrease f nn-assciated prducts. This increase in attentin, culd lead t a better prduct recgnitin as shwn by Pieters et al. (2002). Therefre hypthesis 4 is derived. Hypthesis 4: An increase f recgnitin twards assciated prducts is cupled t a decrease in recgnitin twards unassciated prducts. As explained in sectin 2.2, vlume f music culd als have an effect n viewing behaviur. It is assumed that when peple feel pleasant (like in the tp f the inverted U), peple are mre pen fr assciatins. In this inverted U sft vlume will result in mre attentin cmpared t lud vlume. The increase f the assciatin will result in mre attentin t the assciated prducts. This attentin, can be measured by prduct recgnitin as shwn by Pieters et al. (2002). Therefre hypthesis 5 is derived. Hypthesis 5: The effect f vlume n recgnitin f prducts (assciated with the music) will have the frm f an inverted U. Als in this case, thery frm Pieters and Wedel (2004) culd be applied. It culd be that the mre increase f attentin twards assciated prducts, the mre decrease f attentin twards unassciated prducts. Again, this is measured by prduct recgnitin. Hypthesis 6: The better the recgnitin f assciated prducts, the wrse the recgnitin f unassciated prducts. 15

17 As described in hyptheses 2 and 5 an inverted U curve culd ccur. Hwever, every sng is different. A lw vlume culd be mre suitable fr a certain sng cmpared t anther. Fr example, a jazz sng culd feel pleasant (resulting in mre prduct chices) with a lwer vlume, cmpared t a gspel sng that culd feel mre cmfrtable when played with mre decibels. S the inverted U differs per sng and is nt fixed t a certain amunt f decibels. 16

18 3. Methdlgy The hyptheses that were frmulated in the literature research were tested in a virtual supermarket experiment. Hwever, first was determined which music r sng had a strng assciatin with a prduct r prduct categry. Results f a previus pre-test and pilt study were used. Mrever, a new pre-test was cnducted t find a (secnd) useful sng. Thereafter anther pre-test was dne t determine which vlume per different stage (sft, medium, lud) shuld be used per sng. In the end f this sectin, the methdlgy fr the virtual supermarket is described. In this study all kind f meals culd be mentined, but in the virtual supermarket hwever, several meal kits were used. 3.1 Results f a previus pre-test and pilt study In research dne by Kappen (2014) a pre-test and pilt study were carried ut t measure the assciatin f music with prducts r prduct categries. Amng the different sngs, it became clear that Asian music was mst useful. This type f music had the mst advantages cmpared t the ther sngs. The Asian music had a strng assciatin with the Asian meals Indian Chicken Tandri, Chinese Beef Shanghai and Japanese Teriyaki. Mrever, it had a weak assciatin with all ther (nn-asian) meals. In summary, the tw mst imprtant prs f Asian music were: 1) Music was strngly assciated with Asian meals. 2) Respndents shwed a fast reactin time with the assciatin n riental meals. Therefre, als in the pre-test the same Asian sng was used. The stimulus Asian music was expected t assciate with the prduct categry Asian meals, like Indian Chicken Tandri, Chinese Beef Shanghai and Japanese Teriyaki. Anther pre-test was carried ut t find a secnd useable sng that was assciated with a different meal r nne. New sngs, that were expected t have an assciatin with a certain prduct r prduct categry, were used in the new pre-test. Als tw sngs that were nt expected t have an assciatin with a certain prduct r prduct categry were tested. These sngs culd be used as a cntrl cnditin. 3.2 Pre-test sngs In rder t knw which sng had an assciatin with a prduct r prduct categry, sme tests were dne. Participants In this pre-test 35 respndents participated. All respndents were students. Of the 35 respndents, 18 (51%) were male, 17 (49%) female. The average was years (SD = 1.92), ranging frm 18 t 27 years ld. Stimuli When chsing a certain sng, tw cnditins were taken int accunt. Namely, the sngs did nt mentin the brand, prduct r prduct categry in the sng itself and all the sng culd be 17

19 cnsidered as backgrund music and nt a single (shrt) jingle r tune. Vcal (sng) as well as instrumental sngs were used and are shwn belw. Of every sng, a YuTube link can be fund in appendix 1. Mediterranean (Music that was expected t assciate with Mediterranean meals). Instrumental: Italian Medley (Sng 1). Vcal: Teddy Ren - Picclissima Serenata (Sng 2). Vcal: Claudi Villa Qui Stt Il Ciep Di Capri (Sng 3). Vcal: Dmenic Mdugn Vlare (Sng 4). Instrumental: Sunset in Capri (Tarzia/Ciplla) (Sng 5). Asian (Music that was expected t assciate with Asian meals). Instrumental: Dansngs Ambershire. This sng was already used in a study by Kappen (2014) and had an assciatin with Asian meals (Sng 6). Unassciated music (Music that was nt expected t assciate with a certain prduct r prduct categry). Vcal: Peter Nrth - Slw lve (Sng 7). Instrumental: Backgrund Instrumental Music (Sng 8). Prcedure In the questinnaire, the rder f sngs was mixed. This in rder t make sure that a persn nt listened t 5 sngs f the same expected assciatin after each ther. The playlist fr all respndents is shwn in table 2. Sng number Sng 1 Mediterranean: Italian Medley. Sng 2 Mediterranean: Teddy Ren - Picclissima Serenata. Sng 3 Mediterranean: Claudi Villa Qui Stt Il Ciep Di Capri. Sng 4 Mediterranean: Dmenic Mdugn Vlare. Sng 5 Mediterranean: Sunset in Capri (Tarzia/Ciplla). Sng 6 Asian: Instrumental: Dansngs Ambershire. Sng 7 Unassciated: Peter Nrth - Slw lve. Sng 8 Unassciated: Backgrund Instrumental Music. Table 2: Order f playing the eight sngs f pre-test 1. Played in Order f playing in questinnaire questinnaire # 1 Sng 1 Mediterranean: Italian Medley. # 3 Sng 7 Unassciated: Peter Nrth - Slw lve. # 5 Sng 2 Mediterranean: Teddy Ren - Picclissima Serenata. # 7 Sng 6 Asian: Instrumental: Dansngs Ambershire. # 8 Sng 3 Mediterranean: Claudi Villa Qui Stt Il Ciep Di Capri. # 4 Sng 8 Unassciated: Backgrund Instrumental Music. # 2 Sng 4 Mediterranean: Dmenic Mdugn Vlare. # 6 Sng 5 Mediterranean: Sunset in Capri (Tarzia/Ciplla). 18

20 On befrehand, respndents were asked t listen t a sng first. Frm every sng, 30 secnds f music was heard. This in rder t have an idea f the cmplete music fragment. After listening, respndents had t answer pen questins abut their thughts and whether they culd match a prduct r prduct categry that culd fit the music. Next t that, it was als asked which cuntries were assciated with the sng. Als questins abut arusal, familiarity, liking and the apprpriateness f the sng in a supermarket were asked. Measures As explained, questins were asked abut the assciatin f meal(s) with a certain sng and the assciatin with a cuntry r cuntries with a certain sng. These questins were in an pen frm. Respndents culd answer with a maximum f three meals and/r cuntries, but were nt frced t give an answer. Questins cncerning liking, arusal and apprpriateness f the sng in a supermarket were asked using a slide bar with a scale ranging frm 0 t 100. Familiarity was asked in a clsed frm with just the answers yes r n (fr the whle questinnaire see appendix 2). Belw, the imprtance f every musical aspect is shrtly explained. Strebele & De Castr (2004) fund that music can have an effect n arusal. Arusal can be described as a state in which peple are awake and reactive t stimuli (Cahill & McGaugh, 1998). An increase in arusal can in turn, result in an increase in fd intake and fluid cnsumptin. Next t that, Strebele & De Castr (2004) als stated that when the arusal f a cnsumer increases, it will lead t a greater desire t return t the same spt. This arusal factr shuld be taken int accunt when ding research. Articles have shwn that familiar music results in a mre psitive feeling amng cnsumers. This psitive feeling influences cnsumers. Namely, cnsumers shp lnger when they are expsed t familiar music (Yalch & Spangenberg, 2000; Garlin & Owen, 2006). Besides, shpping expenditure seemed t increase accrding t the level f liking the music (Herringtn & Capella, 1996). S, since familiarity as well as preferences culd have an influence n music and cnsumer behaviur, this shuld be taken int accunt when ding research. Radcy and Byle (1997) suggested that peple spend mre mney and time in a stre r restaurant in case the music played was cnsidered as apprpriate (fit in cntext). Similarly, MacInnis and Park (1991), discussed that peple are easier t persuade in case the music is apprpriate fr the cntext. Anther factr that culd play a rle is temp. Fr example, the genre happy hard-cre will prbably have a higher temp cmpared t classical music. Research has shwn that a slwer temp can lead t a lnger stay f a cnsumer in a retail envirnment, leading t relatively mre purchases cmpared t the extra time stayed (Smith & Curnw, 1966; Caldwell & Hibbert, 2002; Garlin & Owen, 2006). Therefre an expert was asked t search fr music with cmparable, r at least nt significantly different, level f temp. 19

21 Data analysis Answers f the respndents were cmpared. First, a cding scheme is presented belw. As shwn in table 3, the American kitchen was added, since it became clear that a certain sng was assciated with thse meals (hamburger and htdgs). Meals Asian meals Mediterranean meals American meals Others Cuntries Asian cuntries Mediterranean cuntries Others Including Rice, curry, sushi, stir fry dishes, spring rlls and (flur) ndles. Lasagna, pasta, cannli, pizza and paella. Hamburger and htdgs. All ther meals. Japan, China, India, Vietnam, Taiwan, Malaysia, Indnesia and Japan. Spain, Greece, Prtugal and Italy. Cuntries that were mentined less than 2 times (fr ne sng). Table 3: Cding scheme. T cmpare the answers, a cntent analysis was dne cncerning assciatin with meals. When a respndent named three different meals frm ne regin (like Asian meals), this was just cunted as ne assciatin. Fr example, when a respndent was asked what meals were assciated with a certain sng and the respndent answered with lasagna, pizza and pasta, this was cnsidered as ne assciatin with Mediterranean meals (and nt three). This was chsen because the aim was t measure an assciatin with a brad grup f respndents and nt just a cuple f respndent with a very strng assciatin. Respndents culd maximally mentin 3 meals but were nt frced t give an answer. Cncerning the amunt f assciated cuntries, the same methd was used t analyze the results. The nly aspect added, was that a cuntry shuld be mentined at least 2 times (fr ne sng) t be part f the figure, therwise it wuld be mentined as thers. The strength f an assciatin was based n the amunt f respndents that assciate the music with a certain meal r regin. The sngs with the strngest assciatin were used in the virtual supermarket. Results The main gal f this pre-test was t test whether a certain sng was assciated with meals frm a regin. Therefre results shwing assciatins f the 8 sngs with meals were analyzed first. Sng 1 up t and including sng 5 were expected t assciate with Mediterranean meals, sng 6 with Asian meals and sng 7 and 8 t nt assciate with meals. First the sngs cncerning Asian meals are discussed. 20

22 # Assciatins Asian Sng 6 was the nly Asian sng and assciates with meals frm an Asian regin. This sng seems t be very useful, als in this test this sng had a lt f assciatins with Asian meals and nt many with ther meals than Asian. Mediterranean Sng 1 t 5 were expected t have assciatins with Mediterranean meals. Lking at figure 4, it becmes clear that sng 1, 2, and 4 had strng assciatins. Next t that all three sngs did nt have many assciatins with ther meals, just as the Asian sng 6. Therefre sngs 1, 2 and 4 seem t be useful. Unassciated Cncerning the unassciated sngs (7 and 8), it seems that sng 7 had sme assciatins with American meals. Therefre this regin was added. Sng 8 des nt have many assciatins with meals and therefre sng 8 seems useful Mediterranean Asian American Others Figure 4: Number f assciatin per sng with meals frm a regin. Next t the assciatins with meals frm a certain regin, it culd als be f imprtance with what cuntry a certain sng was assciated. In case a certain cuntry was mentined 2 times fr a sng, it is shwn. Otherwise it is shwn in thers. Asian As shwn in figure 5, Asian sng 6 had a lt f assciatins with Asian cuntries. Next t that, it had nly a few assciatins with ther cuntries. These results cnfirm that sng 6 seems t be useful. Mediterranean Mediterranean sngs 1 t 5 did all have many assciatins with Mediterranean cuntries. Since it was already shwn that sng 1, 2 and 4 seemed t be mst useful, the fcus is n these three sngs. Unassciated Sng 7 was many times assciated with the USA. Sng 8 hwever, was nt much assciated with any cuntry, which culd be seen as a psitive aspect fr a sng in the cntrl cnditin. 21

23 # Assciatins Mediterranean Argentina Marcc Netherlands Asian Cuba Mexic Hngarije France Australia Usa Others Figure 5: Number f assciatin per sng with a cuntry r cuntries. Frm the literature backgrund it became clear that musical fit in a certain cntext, arusal, familiarity and liking culd have effects. Therefre questins were asked abut these aspects (fr the whle questinnaire see appendix 2). Only the mst useful sngs were included in table 4. Asian Sng 6 was the Asian sng that was used in the virtual supermarket. It scred quite lw n the musical fit, arusal and liking aspects. Next t that, the sng was quite unknwn since nly 6 percent knew the sng. Mediterranean Lking at the three useful Mediterranean sngs (1, 2 and 4), sng 1 and 2 shw the least differences with Asian sng 6 cncerning musical fit, arusal and liking. Sng 1 hwever, was just like Asian sng 6 instrumental and was therefre preferable. Unassciated Cmparing sng 8 with sng 1 and 6, musical fit has a very high mean and arusal very lw mean. Anther pre-test was dne in rder t find a useful cntrl cnditin (see sectin 3.3). sng 1 Mediterranean sng 2 Mediterranean sng 4 Mediterranean sng 6 Asian sng 8 NA Musical fit (20.57) (20.99) (22.99) (21.78) (24.54) Arusal (17.92) (16.37) (18.39) (17.75) (15.31) Knws the sng % Liking (25.83) (22.54) (21.97) (22.89) (22.92) Table 4: Means and standard deviatin f musical fit, arusal, familiarity and liking f the eight sngs. 22

24 Differences sng 1 and 6 T measure whether the differences cncerning musical fit, arusal and liking frm sng 1 and 6 (as in table 4) were significant, a paired sample t-test was carried ut. Musical fit: There were n significant differences between sng 1 and 6 cncerning musical fit (t (33)= -1.70, p = 0.10). Arusal: There were n significant differences between sng 1 and 6 cncerning arusal (t (30)= 1.35, p = 0.19). Liking: There were n significant differences between sng 1 and 6 cncerning arusal (t (35)= 1.51, p = 0.14). Differences in assciatin with meals sng 1 and 6 As shwn in figure 5, sng 1 had 21 assciatins with the expected regin and 5 ther assciatins cncerning meals. Sng 6 had 25 assciatins with the expected regin and 3 ther assciatins cncerning meals. A chi-square test shwed that the respndents that listened t a certain sng, did nt significantly differ in chsing (assciated) prducts (χ(2) = 0.25, p =.62). Discussin results pre-test Since sng 1 and 6 had strng assciatin with assciated cuntries as well as meals, these sngs seem useful. Next t that, they had nt many assciatins with ther cuntries and meals. Furthermre, the sngs did nt significantly differ cncerning familiarity, musical fit, arusal and liking. S sng 1 (Italian Medley) and sng 6 (Dansngs Ambershire) are used. It was chsen t carry ut anther pre-test in rder t search fr a sng which culd be used as a cntrl cnditin. This sng shuld scre in between r arund the means f sng 1 and 6 cncerning musical fit, arusal, familiarity and liking. 3.3 Pre-test sngs 2 In rder t knw which sng des nt have an assciatin with a prduct r prduct categry, tests were carried ut. The prcedure and data analysis were the same as in pre-test 1. Participants In this pre-test 33 respndents participated. All respndents were students. Of the 33 respndents, 17 (51%) were male, 16 (49%) female. The average was years (SD = 1.72), ranging frm 19 t 25 years ld. 23

25 # Assciatins Stimuli The backgrund sngs that were used, are shwn belw. In appendix 1 YuTube links are prvided. Sng 1: Guitar backing track A minr/c Majr Sng 2: Gerge Ezra - Listen t the Man Sng 3: Easy Listening Music: Guitar & Sax Sng 4: Happy Acustic Backgrund Music - Shining Thrugh by Alum Sng 5: Musse T - Oh Sng Results Again, the prcedure and data analysis were the same as in pre-test 1. Therefre this thesis cntinues with the results. First, the assciatin with meals were analysed. Lking at figure 6, it seems that sng 2, 3 and sng 5 were the mst useful sngs since thse sngs were nt much assciated with Asian r Mediterranean meals Mediterranean Asian American Others 1 0 sng 1 sng 2 sng 3 sng 4 sng 5 Figure 6: Number f assciatin per sng with meals frm a regin. 24

26 # Assciatins Cncerning the (useful) sngs 2, 3 and 5, it seems that sng 3 and 5 were nt much assciated with cuntries, as shwn in figure 7. Especially sng 3 was nt much assciated. S, sng 3 and 5 seem t be gd sngs fr the cntrl cnditin sng 1 sng 2 sng 3 sng 4 sng 5 Mediterranean Argentina Marcc Netherlands Asian Cuba Mexic Hngarije France Australia Jamaica Usa Figure 7: Number f assciatin per sng with a cuntry r cuntries. Cncerning musical fit, arusal, familiarity and liking the cntrl cnditin sng shuld scre in between r near Mediterranean sng 1 and Asian sng 6. As shwn in table 5, sng 5 clearly scred a lwer mean n musical fit and higher n arusal. On liking sng 5 scred als lwer than sng 1 and 6. Sng 3 hwever, scred in between sng 1 and 6 cncerning liking and familiarity. Next t that, cncerning arusal and musical fit it scred means which were relatively nearby sng 1 and 6. sng 3 sng 5 sng 1 and 6 Musical fit (25.67) (18.90) Arusal (14.73) (12.04) Knws the sng % Liking (24.08) (24.57) Table 5: Means and standard deviatin f musical fit, arusal, familiarity and liking cmpared t sng 1 and 6 f the previus pre-test. Differences sng 3 sng 1 and 6. T measure whether the differences in musical fit, arusal and liking frm sng 3 and sng 1 and 6 (as in table 5) were significant, a paired sample t-test was carried ut. Differences cncerning sng 3 and sng 1. Musical fit: There were n significant differences between sng 3 and 1 cncerning musical fit (t (30)= 1.71, p = 0.10). 25

27 Arusal: There were n significant differences between sng 3 and 1 cncerning arusal(t (28)= 1.50, p = 0.88). Liking: There were n significant differences between sng 3 and 1 cncerning liking (t (30)= -0.58, p = 0.56). Differences cncerning sng 3 and sng 6. Musical fit: There were n significant differences between sng 3 and 6 cncerning musical fit (t (30)= 0.22, p = 0.83). Arusal: There were n significant differences between sng 3 and 6 cncerning arusal(t (29)= 1.19, p = 0.24). Liking: There were n significant differences between sng 3 and 6 cncerning liking (t (30)= 0.83, p = 0.41). S, the sngs did nt significantly differ cncerning musical fit, arusal and liking. Discussin results pre-test Since sng 3 (Easy Listening Music: Guitar & Sax) was nt much assciated with meals and cuntries, this sng was used as a cntrl cnditin. Mrever, it scred in between r near Mediterranean sng 1 and Asian sng 6 cncerning musical fit, arusal, familiarity and liking. 3.4 Pre-test vlume Nw the three sngs that will be used are clear, it is f imprtance t determine what vlume shuld be used. As explained, every sng has its wn characteristics; a certain amunt f decibel culd be mre suitable r preferred fr a certain sng cmpared t anther. Therefre a pre-test was dne t examine per sng what culd be cnsidered as sft medium and lud. Participants and design In this pre-test the 10 respndents participated, all students. Of the 10 respndents, 5 were male (50%), 5 female (50%). The average age was 20.7 years (SD = 2.50), ranging frm 18 t 25 years ld. Stimuli Three sngs were used, namely sng 1: Italian Medley, sng 2: Dansngs Ambershire and sng 3: Easy Listening Music: Guitar & Sax. Prcedure The participant listened t every sng and adjusted the sund level in which ne wuld describe the vlume as sft, medium and lud (fr the full questinnaire see appendix 4). The adjusting f sund level (vlume) was dne by using the slider f vlume n the cmputer. This prcedure was dne, per sng per setting ( sft medium and lud ). The average results were used in the virtual 26

28 supermarket. The amunt f decibels per setting per sng was measured with a decibel meter. This decibel meter was placed in frnt f the respndent (abut half a meter). Results In table 6 belw, the means are shwn. The means did nt differ much per sng. In the brackets, the ranges are shwn and they shwed n clear utliers (fr all the results, see appendix 5). Sng 1 Italian medley lw middle high Decibels 40 (37-44) 53 (49-56) 71 (68-74) Sng 2 Dansngs Ambershire lw middle high Decibels 38 (35-40) 49 (46-51) 72 (70-74) Sng 3 Easy Listening Music: Guitar & sax lw middle high Decibels 38 (35-40) 47 (43-49) 68 (66-73) Table 6: Means f the three sngs, cncerning a lw, middle and high vlume. Discussin results pre-test vlume The means f the decibels per setting per sng, as described in table 6, were used in the virtual supermarket. 3.5 Try-ut Befre the experiment in the virtual supermarket tk place, a try-ut was dne t find ut whether there were any imprvements pssible. Methd 3 Dutch students frm the WUR participated, f which 2 were male (age 19 and 21), 1 female (age 22). The experiment tk apprximately 20 minutes. Because nly three students participated, nly ne f the three sngs was used. In this case was chsen fr the Asian sng (Dansngs Ambershire). A vlume f 49 decibels was used. The virtual supermarket was displayed n three large screens and music was played by speakers. Prcedure The participants participated ne after anther. They read the task descriptin as in appendix 9. Thereafter they used the virtual supermarket and cmpleted the tasks. Afterwards they filled in a questinnaire (as in appendix 10) n a laptp. At last, they were asked whether anything was unclear r culd be imprved. 27

29 Results try-ut Since the aim f the try ut was t find ut whether anything culd imprve, nly results cncerning this aspect were discussed. The participants mentined 4 imprvements, 3 cncerning the virtual supermarket, 1 cncerning the questinnaire. The first aspect was that the participants fund that they had t be t clse t a prduct befre they culd pick it up. The distance, t be able t pick up a prduct, shuld be larger in rder t imprve user-friendliness. The secnd aspect was that the start psitin in the virtual supermarket was quite strange, namely with the eyes pinting at the wall. When a participant wanted t walk int the right directin, a 180 degree turn was directly necessary. The third aspect was that sme meal packages were bigger than thers, while this shuld nt be the case. Cncerning the questinnaire it became clear that the lading f the images tk relatively much time. The 4 nticed pssible imprvements, which became clear, were applied befre the virtual supermarket experiment tk place. S, the distance t be able t pick up a prduct was made larger, the start psitin was changed, the meal packages have the same size and the images have fewer pixels in rder t reduce the lading time. 3.6 Virtual supermarket Participants and design Participants were 294 students frm Wageningen University ranging in age frm 18 t 37 years (M = 20.74, SD = 2.44), with 153 (52%) f them female and 141 (48%) male. During the experiment the male/female percentage per cnditin was taken int accunt in rder t avid majr differences. Students with clur blindness were excluded frm participatin. Participants participated in ne f the cnditins frm a 3 (sng type: Asian, Mediterranean, neutral) x 3 (vlume: lud, medium, sft) between subjects design. The cnditins were equally divided between mrnings and afternns, as shwn in appendix 6. Stimuli Music As befre, three different sngs were used. Namely: Asian sng (Dansngs Ambershire), Mediterranean sng (Italian Medley) and (in the cntrl grup) neutral sng (Easy Listening Music: Guitar & Sax). These different sngs were mixed with backgrund sunds frm a supermarket, e.g. the bleep sund f the pay desk. The backgrund sunds were added in rder t make the virtual supermarket mre realistic. Per cnditin, a different vlume was used. The mean amunt f decibels which was mst suitable fr the cnditin was tested in a pre-test. Table 7 n the next page prvides an verview. 28

30 Music + vlume level Mean db Asian Lud 72 Asian Medium 49 Asian Sft 38 Mediterranean Lud 71 Mediterranean Medium 53 Mediterranean Sft 40 Neutral Lud 68 Neutral Medium 47 Neutral Sft 38 Table 7: Mean amunt f db per cnditin. Prducts When the prducts (that were assciated with certain music) were chsen, the differences between prducts were taken int accunt. Research has namely shwn that when a prduct des nt pssess bjective advantages r differences, cmpared t ther prducts, persuasin with backgrund music is mre effective (Ktler, 1973; Batra & Ray, 1983). A high degree f similarity in mst attributes is called prduct parity (Myers, 2003). S, if prduct attributes are cmparable and d nt pssess lt f differences r advantages (prduct parity), persuasin is easier. Fr example, tw types f meal packages can be bught. When bth prducts are cmparable amng mst attributes, but nly differ in, fr example, rigin. Persuasin with making use f this rigin is easier cmpared t a case in which the prducts differ in mre attributes. Fr example, when prduct a is als strikingly less salty cmpared t prduct b. Prducts that were nudged were meal packages frm different regins. These prducts were cmparable and the meal packages did nt pssess bjective advantages r differences between each ther. In this case Asian and Mediterranean meals were nudged against ther meals. Cncerning the meal packages, per regin (Asian/ Mediterranean/ thers) 10 prducts were placed in the virtual supermarket. In appendix 7 all prducts that were part f this research are shwn and als a screensht f the shelf is shwn. Prcedure The study cnsisted f three phases: (1) intrductin, cnsent and infrmatin (2) ding grceries in a virtual supermarket (3) an additinal questinnaire. When participants walked int the rm in which the experiment tk place, the backgrund music was already playing. The amunt f decibels (f the music) was measured by using a decibel meter, the decibel meter was placed in frnt f the participants seat (abut half a meter) and when the right amunt f decibels was set, participants were welcme and the decibel meter was remved. First, participants filled in the cnsent frm. In this cnsent frm backgrund infrmatin was described and participants signed when they agreed. The cnsent frm is shwn in appendix 8. Thereafter participants tk place in frnt f the virtual supermarket. At this mment they read the task descriptin, as shwn in appendix 9. After reading, the experiment tk place. Participants did grceries with a given shpping list (as shwn in the task descriptin) in a virtual supermarket. In this given shpping list was instructed t buy three prducts (apple, pack f ckies and a package f margarine). This was dne s that participants get used t the virtual supermarket and the 29

31 navigatin, befre a participant bught a meal package. An verview f the virtual supermarket was shwn in the task descriptin in rder t make it easier t find the prducts. After ding the grceries, participants had t fill in a questinnaire n a laptp (fr the full questinnaire, see appendix 10). Questins that were asked were related t their emtins, experiences and prduct chice. Cncerning emtins, the 20 items f the PANAS (Psitive and Negative Affect Schedule) (Watsn et al., 1988) were used, which is cnsidered as a reliable and valid measure (Crawfrd & Henry, 2004). Cncerning experiences, was fr example asked whether they had nticed striking aspects. Cncerning prduct chice, questins cncerning the recgnitin f prducts were asked. Of 21 prducts was asked whether the participants had seen the prducts in the shelf r nt. The rder f questins and whether they were in shelf r nt, is shwn in appendix 11. Per sng type 7 prducts were shwn, f which 5 were in the shelf. Furthermre, participants had t answer backgrund questins like age, gender and what they think the purpse f the study was. Participatin lasted arund 20 minutes and participants received a reward (chclate bar, bag f nuts r bag f chips). Measures Prduct chice Per respndent culd be seen which meal package was chsen and in which cnditin the participant participated. The chsen meal package culd be a Mediterranean, Asian r ther meal package. Prduct picked up Als data cncerning the prducts that were picked up were cllected by the experiment using the virtual supermarket. Per respndent culd be seen whether and hw many times a Mediterranean, Asian and/r ther meal packages was/were picked up. Als in this case was registered in which cnditin the participant participated. Prduct recgnitin Of 21 prducts was asked, in the questinnaire, whether a persn had seen that prduct in the virtual supermarket. The wrng answers cncerning the recgnitin f 21 prducts were cunted per persn. Als the wrng answers cncerning the recgnitin f nly the prducts that were in the shelf (15 prducts) were cunted. Emtins and md Als questins abut emtins and md were asked in the questinnaire. The 20 items f the PANAS were used (10 psitive, 10 negative)(watsn et al., 1988). T find ut whether items within the subscale psitive md items were reliable, Crnbach s Alpha was measured. The subscale cnsisted f the items geïnteresseerd, pgewekt, sterk, enthusiast, zelfverzekerd, alert, geinspireerd, vastberaden, aandachtig and energiek, and its Crnbach s Alpha was.80. This suggesting that the items have relatively high internal cnsistency (Field, 2009). In case any item wuld be deleted Crnbach s Alpha wuld nt increase. 30

32 T find ut whether items within the subscale negative md items were reliable Crnbach s Alpha was measured. The subscale cnsisted f the items bedrefd, terneergeslagen, schuldig, angstig, vijandig, geirriteerd, bang, gespannen, beschaamd and zenuwachtig. Its Crnbach s Alpha was.90. This suggesting that the items have relatively high internal cnsistency (Field, 2009). In case any item wuld be deleted Crnbach s Alpha will nt significantly increase. S, 2 new variables were created, which were the mean f psitive items (named Psitive items) and the mean f the negative items (named Negative items ). Experiences and pinin Cncerning experiences and pinin, was asked fr example whether the participant nticed striking aspects and whether the situatin was realistic. Als questin were asked abut their fd cnsumptin. Fr example, if the participants (r whether they have taken int accunt thers that) d nt eat particular fd, and if they nrmally wuld use a meal package. Als questins abut the liking f music were asked, namely: He (n)prettig vnd u de muziek? and He (n)aangenaam vnd u de muziek?.t find ut whether the items were reliable, Crnbach s Alpha was measured. The subscale cnsisted f the items (n)prettigheid and (n)aangenaamheid, and its Crnbach s Alpha was.83. This suggesting that the items have relatively high internal cnsistency (Field,2009). S, a new variable was created, which was named liking f music. Als questins with regard t the fit f the music in a supermarket envirnment were asked, namely: He passend (in een supermarkt mgeving) vnd u de muziek? and He geschikt (vr in een supermarkt mgeving) vnd u de muziek?.t find ut whether the items were reliable Crnbachs Alpha was measured. The subscale cnsisted f the items passendheid and geschiktheid, and its Crnbachs Alpha was.86. This suggesting that the items have relatively high internal cnsistency (Field, 2009). S, a new variable was created, which was named music fit. Backgrund characteristics Finally, backgrund characteristics f the participants were cllected. These include sci demgraphic characteristics, like age and gender. 31

33 4. Results virtual supermarket 4.1 Statistical analysis Fr the results, shwn in chapter 4, the prgram SPSS was used (versin number 22). In the results chi-square tests and tw-way ANOVA tests were carried ut. Fr hypthesis 4 crrelatin tests were carried ut and fr hypthesis 5 als a repeated measures ANOVA was cnducted. 4.2 Sample All students spke and understd the Dutch language. Next t the Dutch natinality, als 7 students with a German natinality participated. N prblems cncerning this natinality were expected because they stated that they understd and spke the Dutch language. Next t that, these German students had been living at least 7 mnths in The Netherlands. 4.3 Similarity cnstrains The Asian, Mediterranean and cntrl grup shuld be similar cncerning t ther aspects that were f imprtance in this research. Therefre these grups were cmpared n several aspects by making use f the results f the questinnaire (as in appendix 10). In this paragraph, these results are described. Age Althugh the average age between the types f sng did nt seem t differ much (Asian M =21.28, Mediterranean M = 20.37, Neutral M = 20.55) there was a significant main effect f type sng n age (F(2, 294) = 3.83, p =.02). Als cncerning the means f the vlumes (Lud M = 20.29, Medium M = 20.75, Sft M = 21.16), the differences seemed small. Hwever, there was als a significant main effect f vlume n age (F(2, 294) = 3.17, p =.04), as well as the type sng x vlume interactin (F(4, 294) = 2.58, p =.04). Table 8, belw, prvides an verview f the mean f the age per cnditin. Sng heard Vlume f N Mean SD music Asian Lud Medium Sft Ttal Mediterranean Lud Medium Sft Ttal Neutral Lud Medium Sft Ttal Table 8: Overview f the average age per cnditin. 32

34 The differences in age culd influence this research, therefre these differences were taken int accunt when analysing. Gender The participants that listened t a certain sng type, did nt differ in gender (χ(2) = 0.19, p =.91). Als the participants that heard a certain vlume, did nt differ in gender (χ(2) = 0.97, p =.62). Table 9 prvides an verview f the participants per cnditin and shws the number f male/female. In ttal 153 participants were female (52%) and 141 male (48%). Asian Sng Asian Lud # % Asian Medium # % Asian Sft # % Male 15 47% Male 18 56% Male 16 44% Female 17 53% Female 14 44% Female 19 54% Ttal 32 Ttal 32 Ttal 35 Mediterranean Sng Mediterranean # % Mediterranean # % Mediterranean # % Lud Medium Sft Male 17 52% Male 14 44% Male 14 44% Female 16 48% Female 18 56% Female 18 56% Ttal Ttal 32 Neutral Sng Neutral Lud # % Neutral Medium # % Neutral Sft # % Male 17 53% Male 16 48% Male 14 42% Female 15 47% Female 17 52% Female 19 58% Ttal 32 Ttal 33 Ttal 33 Table 9: Participants per cnditin. S, there was n difference between male and female regarding type f sng heard and vlume. Thus, pssible effects cannt be attributed t differences in gender. Buying meal packages Cncerning the frequency f buying meal packages, 14.6 percent f the participants stated that they never use a meal package. The table cncerning whether and hw frequent participants use a meal package is shwn belw. Answer Frequency Percent N, I never use meal packages Yes, less than nce a year Yes, 1 time in the 6 mnths t nce a year Yes, 1 time in the 3 t 6 mnths Yes, 1 time in the mnth t 3 mnths Yes, 1 time in the tw weeks t 1 mnth Yes, 1 time in the week t tw weeks Yes, 2 t 5 times a week Ttal Table 10: Frequency f using a meal package. 33

35 The participants that listened t a certain sng type, did nt differ in the frequency f buying meal packages (χ(14) = 20.44, p =.12). Als the participants that heard a certain vlume, did nt differ in the frequency f buying meal packages (χ(14) = 12.35, p =.58). The participants that listened t a certain sng type and vlume, did nt differ in the frequency f buying meal packages. Thus, pssible effects cannt be attributed t differences in buying meal packages. Music familiarity Table 11 belw prvides an verview with regard t familiarity f the sngs. Lking at the percentages, n majr differences are fund. In the three different sng cnditins, the mst f the participants were nt familiar with the sng. Only sme f them stated that they heard the sng, but did nt knw the title r artist. Nne f the participants knew the artist and/r title f the sng. The participants that listened t a certain sng type, did nt differ in familiarity with the sng (χ(2) = 2.70, p =.26). Als the participants that heard a certain vlume, did nt differ in familiarity with the sng (χ(2) = 5.29, p =.77). Was u bekend met het nummer? Muziek gehrd Asian Mediterranean Neutral Ttal Nee 89% 83% 91% 87% Ja, ik het nummer wel eens gehrd, maar 11% 17% 9% 13% ken titel en artiest niet Table 11: Familiarity f the sngs. The participants that listened t a certain sng type and vlume, did nt differ in familiarity with the sng. Thus, pssible effects cannt be attributed t differences in familiarity with the sng. Music liking There was a significant main effect f type sng n music liking (F(2, 267) = 4.10, p =.02). N effect f vlume n music liking was fund: F(2, 267) = 0.55, p =.58. N interactin was fund between type sng x vlume n music liking (F(4, 267) = 1.01, p =.40). The significant effect f type sng n music liking was further explred with a pst hc test (using LSD). This test shwed that the nly significant difference was the liking f Asian and neutral music (p =.01). The means shw that the neutral music was mst liked (M = 63.40, SD = 19.42), fllwed by the Mediterranean music (M = 58.55, SD = 20.82, and Asian as least liked (M = 54.35, SD = 24.69). Unfrtunately, there was a significant difference cncerning music liking and type sng. Hwever, the partial η² =.03, which means that this measurement f effect size shws that this was a small effect (Field, 2009). The differences in liking f the music culd influence this research, therefre these differences were taken int accunt when analysing. Music fit There was a significant main effect f type sng n music fit (F(2, 267) = 14.14, p <.001). N effect f vlume n music fit was fund: F(2, 267) = 0.70, p =.50. N interactin was fund between type 34

36 sng x vlume n music fit (F(4, 267) = 1.60, p =.18).The significant effect f type sng n music fit was further explred with a pst hc test (using LSD). This test shwed that there is a significant difference between neutral and Asian (p <.001), neutral and Mediterranean (p <.01) and between Mediterranean and Asian (p =.03). The means shw that neutral music was cnsidered as mst apprpriate (M = 52.31, SD =23.56), fllwed by the Mediterranean music (M = 41.95, SD = 24.94) and lastly the Asian sng (M = 34.83, SD= 21.15). S, unfrtunately, there was a significant difference cncerning music fit and type sng. Hwever, Partial η² =.1, which means that this measurement f effect size shws that this was a weak effect (Field, 2009). The differences in musical fit culd influence this research, therefre these differences were taken int accunt when analysing. Md There was n effect f type sng n psitive items fund (F(2, 294) = 0.77, p =.46). Als n effect f vlume n psitive items was fund (F(2, 294) = 0.25, p =.78). N interactin was fund between type sng x vlume n psitive items (F(4, 294) = 0.88, p =.48). Overall, participants psitive md items scred a mean f (SD = 0.60). There was als n effect f type sng n negative items fund (F(2, 294) = 0.15, p =.86). Als n effect f vlume n negative items was fund (F(2, 294) = 2.26, p =.11). N interactin was fund between type sng x vlume n negative items (F(4, 294) = 0.66, p =.62). Overall, participants negative md items scred a mean f (SD = 0.82). T be sure, each item was cmpared individually, and still n significant differences were fund (fr details, see appendix 12). In ttal it can be cncluded that participants md in the cnditins was similar. Overall cnclusin fr similarity cnstrains : There are n significant differences between the cnditins with regard t gender, frequency f buying meal packages, familiarity f music and md. Effects were fund with regard t age, music fit and music pinin. These effects were small, as shwn by the partial η² and Cramer s V (Field, 2009). S, nly small differences exist. Therefre, all the participants included and grups f participants were cnsidered as equal. The significant differences in age, music liking and music fit, were taken int accunt. This was dne by carrying all tests f this experiment ut with age, music fit and music liking as a c-variant. N significant effects f age, music liking and music fit were fund. Next t that age, music liking and music fit had n majr influence n utcmes f the virtual supermarket experiment. Fr example, it did nt influence whether an utcme was significant r nt. 35

37 Means f ludness f music (accrding t participants) 4.4 Manipulatin check Opinin f vlume f music heard There was n effect f type sng n pinin f vlume fund (F(2, 267) = 2.01, p =.14). There was a significant main effect f vlume n pinin f vlume : F(2, 267) = 46.69, p <.001). N interactin was fund between type sng x vlume n pinin f vlume (F(4, 267) = 1.01, p =.40). The significant effect f vlume n vlume pinin was further explred with a pst hc test (using LSD). This test shwed that there was a significant difference between lud and sft (p <.001), between lud and medium (p = <.001) and between medium and sft (p <.001). The means shw that lud music is cnsidered as mst lud (M = 59.59, SD = 11.54), fllwed by the medium music (M = 53.69, SD = 10.31) and sft music lastly (M = 44.22, SD= 10.93). The significant main effect f vlume n pinin f vlume was expected. It was intended that a lud vlume was cnsidered as mre ludly, cmpared t medium r sft vlume by the participants. Lking at figure 8, it shws that the expectatins are met SD Mean 10 0 Lud Medium Sft Vlume f music heard Figure: 8: Ludness f music per vlume cnditin. S, the pinin f vlume was as expected (lud is cnsidered as lud, medium as medium, sft as sft). Furthermre, there was n effect f type sng n pinin f vlume. This is a psitive result, because effects f vlume cannt be attributed t different types f sngs. Music heard Althugh music was played during all experiments, it might be that sme participants did nt hear the music. The participants that listened t a certain sng type, differed in hearing the music during the experiment. The type f sng heard had an effect n hearing the music during the experiment (χ(2) = 6.88, p =.03). The participants that heard a certain vlume, did nt differ in hearing the music during the experiment (χ(2) = 3.15, p =.21). S vlume did nt influence whether a sng was heard. 36

38 The Asian music was heard mre times than the neutral music (97% cmpared t 87%) and als the Mediterranean music was heard a little bit mre times cmpared t the neutral music (89% cmpared t 87%) Table 12 prvides an verview cncerning the amunt f participants that heard music per cnditin. Vlume f music Music heard Sng type Asian Mediterranean Neutral Lud N 0% 12% 3% Yes 100% 88% 97% Medium N 9% 16% 12% Yes 91% 84% 88% Sft N 0% 6% 24% Yes 100% 94% 76% Ttal N 3% 11% 13% Yes 97% 89% 87% Table 12: Frequency f music heard per cnditin. A pssible explanatin fr the difference cncerning music heard and type sng is that the Asian and als, althugh t a lesser extent, Mediterranean music was mre striking than the music frm the cntrl grup, because peple were nt used t this type f music in a supermarket envirnment. The Cramer s V is 0.15, therefre this result can be cnsidered as having nly a small effect size (Field, 2009). The differences in hearing the music during the experiment culd influence this research, therefre these differences were taken int accunt when analysing. Realism f virtual supermarket Of the 294 participants, 98 (33%) stated smething abut ne r mre unrealistic aspect(s) f the virtual supermarket. These 98 answers were fund by a cntent analysis cncerning the questin: What did yu remember f the virtual supermarket? Did yu nticed smething abut the supermarket?. The mst heard cmments cncerning the realism, were that the supermarket was small, had a lt f grey clur with n decratin (like psters r advertisements) and that there was n cash desk. The pinin f the realism f the virtual supermarket did nt differ between the cnditins. The type f sng heard had n effect n the pinin f the realism f the virtual supermarket (χ(2) = 2.01, p =.37). The participants that heard a certain vlume, did nt differ in the pinin f the realism f the virtual supermarket (χ(2) = 3.03, p =.22). 37

39 4.5 Results Prduct chice The participants that listened t a certain type f sng, did nt significantly differ in chsing prducts (χ(4) = 3.03, p =.55). In table 13 ne can see the distributin f the amunt f prducts that were chsen per cnditin. Type f sng + vlume Prduct Chsen Ttal Asian Mediterranean Other # % # % # % # Asian Lud Asian Medium Asian Sft Ttal Asian Mediterranean Lud Mediterranean Medium Mediterranean Sft Ttal Mediterranean Neutral Lud Neutral Medium Neutral Sft Ttal Neutral Table 13: Prduct purchases per cnditin. Als the participants that listened t a certain vlume level, did nt significantly differ in chsing prducts (χ(4) = 0.82, p =.94). Als ther tests were carried ut, in which participants that listened t the same type f sng, but a different vlume were cmpared. Cncerning Asian music, with different vlumes, the difference with regard t chsen prducts was nt significant (χ(4)= 0.86, p =.93.) Cncerning Mediterranean music, with different vlumes, the difference with regard t chsen prducts was nt significant (χ(4)= 1.40, p =.85). Cncerning neutral music, with different vlumes, the difference with regard t chsen prducts was nt significant (χ(4) =3.84, p =.43). Table 14 n the next page, prvides a clear verview. The percentages f the chsen prducts per different vlume level, did nt have the frm f an inverted U. S, a medium level f assciated music, did nt result in mre chsen prducts (f assciated prducts) cmpared t a lud r sft vlume. 38

40 Sng Heard Prduct Chsen in % Asian Asian Lud 44 Asian Medium 34 Asian Sft 34 Mediterranean Mediterranean Lud 52 Mediterranean Medium 47 Mediterranean Sft 41 Neutral Neutral Lud 24 Neutral Medium 21 Neutral Sft 29 Table 14: Relevant prduct chsen per cnditin. In cnclusin, there seemed t be n effect f vlume r type f sng n purchase f prducts (assciated with the music). Thus, it did nt matter which type f sng r vlume level the sng was played with regard t which prduct participants chse. S, hypthesis 1 as well as hypthesis 2 are rejected. H1: When cnsumers listen t music that is assciated with a prduct r prduct categry, this will lead t mre prduct purchases f that prduct r prduct categry cmpared t when cnsumers listen t music that is nt assciated. H2: The effect f vlume n purchase f prducts (assciated with the music) will have the frm f an inverted U. Prduct picked up Next t the chsen prducts, als data with regard t whether participants picked up a prduct (Asian, Mediterranean r ther) were analyzed. As described in 2.3 this attentin is f imprtance since prducts which are prcessed in wrking memry (due t attentin) will stand ut later n. This standing ut, culd result in mre attentin which, in turn, might result in a purchase. The participants that listened t a certain type f sng, did nt significantly differ in picking up prducts. The difference cncerning Mediterranean prduct(s) picked up was nt significant (χ(2) = 0.26, p =.88). The difference cncerning Asian prduct(s) picked up was nt significant (χ(2) = 3.15, p =.21). The difference cncerning ther prduct(s) picked up was als nt significant (χ(2) = 1.65, p =.44). Table 15 n the next page, shws the results cncerning the distributin f the number f participants that viewed a certain type f prduct (Asian, Mediterranean r ther) per cnditin. 39

41 Sng Heard Prduct Picked Up Ttal Asian Mediterranean Other # % # % # % # Asian Lud 14 44% 13 41% 5 16% 32 Asian Medium 12 33% 18 50% 6 17% 36 Asian Sft 12 33% 17 47% 7 19% 36 Ttal Asian 38 37% 48 46% 18 17% 104 Mediterranean Lud 10 28% 18 50% 8 22% 36 Mediterranean Medium 10 29% 16 47% 8 24% 34 Mediterranean Sft 12 36% 14 42% 7 21% 33 Ttal Mediterranean 32 31% 48 47% 23 22% 103 Neutral Lud 10 30% 15 45% 8 24% 33 Neutral Medium 11 32% 16 47% 7 21% 34 Neutral Sft 5 14% 20 57% 10 29% 35 Ttal Neutral 26 25% 51 50% 25 25% 102 Table 15: Prduct picked up per cnditin. Als tests were carried ut in which nt nly was analysed whether a certain prduct type (Asian, Mediterranean, ther) was picked up, but als the number f times a certain prduct type was picked up. In this test the effect f sng type n ttal Asian prducts picked up was examined. The same tests were carried ut cncerning Mediterranean and neutral prducts picked up. Again, n significant differences were fund (Asian: F(2, 293) = 1.18, p =.31 Mediterranean: F(2, 293) = 0.54, p =.58 and neutral: F(2, 293) = 2.32, p =.10). Als, the participants that listened t a certain level f vlume, did nt significantly differ in picking up prducts. The difference cncerning Mediterranean prduct(s) picked up was nt significant (χ(2) = 0.39, p =.82). The difference cncerning Asian prduct(s) picked up was nt significant (χ(2) = 0.94, p =.62). The difference cncerning ther prduct(s) picked up was als nt significant (χ(2) = 0.21, p =.90). Als when tests were carried ut cncerning the number f times a certain prduct type was picked up, n significant effect f vlume were fund. In this test the effect f vlume n ttal Asian prducts picked up was examined. The same tests were carried ut cncerning Mediterranean and neutral prducts picked up. Again, n significant differences were fund (Asian: F(2, 293) = 0.01, p =.99 Mediterranean: F(2, 293) = 0.40, p =.96 and neutral: F(2, 293) = 0.11, p =.89). In cnclusin, there seemed t be n difference between cnditins with regard t the meal package picked up. Thus, it did nt matter which vlume the music r which type f sng was played with regard t which prduct(s) participants picked up. S, this enhances the idea that hypthesis 1 as well as hypthesis 2 can be rejected. 40

42 Prduct recgnitin There was n effect f type sng n wrng answers fund. Regarding wrng answers cncerning Asian prducts n significant effect was fund (F(2, 293)= 0.98, p =.38), regarding wrng answers cncerning Mediterranean prducts n significant effect was fund (F(2, 293)= 0.05, p =.96), and als regarding ther prducts n significant effect was fund (F(2, 293)= 0.48, p =.62). Table 16 belw, shws the means f wrng answer per cnditin, per prduct type (Asian, Mediterranean r ther prducts). Music heard Als when the wrng answers f the recgnitin f the prducts that were in the shelf were analyzed, n significant effect f type sng n wrng answers was fund (Wrng answers cncerning Asian prducts: F(2, 293)= 1.81, p =.16, Mediterranean prducts: F(2, 293)= 0.15, p =.86, ther prducts: F(2, 293)= 0.07, p =.94). Fr each type f sng (Asian, Mediterranean r neutral), tests with a different vlume were cmpared. Cncerning Asian music, with different vlumes, the difference with regard t recgnitin f Asian prducts was nt significant (F(2, 99)= 0.75, p =.47). Cncerning Mediterranean music, with different vlumes, the difference with regard t recgnitin f Mediterranean prducts was nt significant (F(2, 97)= 0.09, p =.92). Cncerning neutral music, with different vlumes, the difference with regard t recgnitin f ther prducts was nt significant (F(2, 98)= 0.21, p =.81). In table 17 the means f wrng answers per sng heard and vlume cnditin are shwn. An inverted U frm is nt visible. Sng heard Vlume Mean f wrng answers (SD) Asian Mediterranean Neutral Wrng cncerning Asian prducts Means f wrng answer(s) (SD) Wrng cncerning Mediterranean prducts Wrng cncerning ther prducts Asian 2.12 (1.67) 2.30 (1.41) 3.04 (1.80) Mediterranean 2.38 (1.46) 2.35 (1.42) 2.80 (1.88) Neutral 2.41 (1.63) 2.30 (1.24) 2.83 (1.91) Table 16: Means f wrng answer(s) per cnditin. Table 17: Means f wrng answer(s) per sng ad vlume level. Cncerning Asian Prducts Hard 1.97 (1.43) Medium 1.97 (1.73) Sft 2.40 (1.82) Cncerning Mediterranean Prducts Hard 2.30 (1.36) Medium 2.44 (1.50) Sft 2.31 (1.45) Cncerning Other Prducts Hard 2.84 (1.85) Medium 2.97 (2.07) Sft 2.67 (1.85) 41

43 In cnclusin, there seemed t be n difference between cnditins with regard t prduct recgnitin. Thus it did nt matter which type f sng r vlume f a sng was played cncerning recgnitin f prducts. S, hypthesis 3 and 5 are rejected. H3: When cnsumers listen t music that is assciated with a prduct r prduct categry, these prducts r prduct categry will be better recgnized cmpared t when cnsumers listen t music that is nt assciated H5: The effect f vlume n recgnitin f prducts (assciated with the music) will have the frm f an inverted U. A repeated measures ANOVA was cnducted t cmpare the effect f sng type n the wrng answers f prduct recgnitin cncerning Asian, Mediterranean and ther prducts. The test shwed that there was n significant interactin effect f sng type (F(2, 291)= 0.22, p =.80). In cnclusin, an increase f recgnitin twards assciated prducts was nt cupled t a decrease in recgnitin twards unassciated prducts. Since hypthesis 3 is rejected, it might nt be surprising that als hypthesis 4 is rejected. H4: An increase f recgnitin twards assciated prducts is cupled t a decrease in recgnitin twards unassciated prducts. Crrelatin tests shw that wrng answer Asian, wrng answer Mediterranean and wrng answer thers were significantly crrelated. The wrng answers cncerning Asian prducts significantly crrelated with wrng answers cncerning Mediterranean prducts (r(292) =.13, p =.03). The wrng answers cncerning Asian prducts crrelated significantly with wrng answers cncerning thers prducts (r(292) =.23, p <.001). The wrng answers cncerning Mediterranean prducts crrelated significantly with wrng answers cncerning thers prducts (r(292) =.14, p =.02). These crrelatins might indicate that the better the recgnitin f assciated prducts, the better the recgnitin f unassciated prducts. Therefre these crrelatins were analyzed in detail. Table 18 belw prvides an verview f the crrelatins cncerning wrng answer Asian, wrng answer Mediterranean and wrng answer thers when a certain sng type was heard. Althugh nt all crrelatins were statistically significant, the Pearsn s r was in all the results psitive. This indicates that sme participants were relatively gd in recgnitin f prducts and thers had mre difficulties in recgnitin. This is nt in line with the hypthesis, which presuppses a cntrast f recgnitin between assciated prducts and unassciated prducts. Sng Type N Wrng Asian Wrng Asian Wrng Med Wrng Med Wrng Others Wrng Others r p r p r p Asian Mediterranean Neutral Table 18: Crrelatin in case a certain type f sng is heard. 42

44 Since this hypthesis is in the extensin f hypthesis 4, it is nt surprising that this hypthesis is nt accepted. In cnclusin, the better the recgnitin f assciated prducts, the better the recgnitin f unassciated prducts. S, hypthesis 6 is rejected. H6: The better the recgnitin f assciated prducts, the wrse the recgnitin f unassciated prducts. 43

45 5 Cnclusin 5.1 Discussin The aim f this study was t find ut what the effect is f a strngly assciated sng, cmbined with a certain vlume, n cnsumers prduct chice (in a virtual supermarket). Frm the pre-tests was cncluded that the Asian sng (Dansngs Ambershire), Mediterranean sng (Italian Medley) and neutral sng (Easy Listening Music: Guitar & Sax) were mst useful. These sngs were used in the virtual supermarket experiment. N significant results f the sngs n prduct chice and prduct recgnitin were fund. Besides, vlume f music did nt have an influence n bth prcesses (prduct chice and prduct recgnitin). In cntrast t results f Agmn (1990), Areni & Kim (1993) and Nrth et al. (1993), this study fund that music had n significant effect n prduct chice. There are several ways which culd explain this (nn) effect. First f all, this result may be due t the (in) apprpriateness f the Mediterranean and Asian music in a supermarket envirnment. As shwn in the results, bth sngs scred relatively lw n the apprpriate aspect. Wilsn (2003) shwed that peple cnsumed mre beverages in a restaurant when jazz, easy listening r ppular music was played, cmpared t classical music due t the apprpriateness. In additin MacInnis and Park (1991), discussed that peple are easier t persuade in case the music is apprpriate fr the cntext. Althugh nudging is nt the same as persuasin, it might be that n significant results were fund because the music was cnsidered as t inapprpriate. Anther reasn might be that cncerning meal packages participants had a strng pinin. It culd be that ther prducts r ther prduct categries are easier t nudge. This might als be a reasn why Areni & Kim (1993) and Nrth et al. (1993) fund significant effects f music n prduct chice. Since they used wine as a prduct categry, which might be a prduct where peple have a less strng pinin abut. Besides, n significant effects f music n prduct recgnitin were fund. This might be due t that peple s mind-set was nt fcused n remembering the specific prducts. It was nt their task, and there was n need t remember these prducts. It might be that string aspects in memry is a mre cnscius prcess. In fact, Ynelinas (2001) fund that cnsciusness is ne f the three crucial c s (next t cntrl and cnfidence) fr recgnitin. S, because peple were nt cnsciusly busy with string the prducts which they had seen, it might be that n significant effect were fund. Als n effect f vlume n bth prcesses (prduct chice, prduct recgnitin) was fund. S, it seems that awareness f the music did nt influence the musical assciatin n specific prducts. It seems likely that the vlume did nt influence the prcesses (prduct chice, prduct recgnitin) because the manipulatin was well applied. It became clear that peple cnsidered lud music as lud, medium as medium and sft as sft. Anther reasn fr finding n effect might be that the differences between sft, medium and lud were nt distinctive enugh t find significant effects, this in spite f earlier pre-tests with vlume. In case the three vlume cnditins were nt distinctive enugh, thse cnditins are seen as equal and n effect f different vlumes was measured. 44

46 In cnclusin, this research seems nt t supprt the thery that strngly assciated music can influence prduct chice and recgnitin (memry). Furthermre was fund that vlume des nt have an influence in this prcess. 5.2 Limitatins and future research Althugh the Asian, Mediterranean and neutral sng in the pre-test had n significant differences in liking r apprpriateness in the virtual supermarket experiment significant differences were fund. This might be the case because in the pre-tests participants listened t a lt f sngs, ne after anther. This fixed rder f playing the sngs, might have influenced the answers. Fr example, ne might rate a sng higher cncerning liking, because the sng befre was definitely nt what the participant liked. Hwever, when participants participated in the virtual supermarket, this was nt the case and they listened t nly ne sng. Anther reasn might that the number f participants in the pre-test (respectively 35 and 33) were t small. It might be that when mre respndents participated, the results were als significant difference. In future research a randmized rder f playing the sngs (in the pre-tests) is a slutin fr this limitatin. Besides, mre participants in the pre-tests wuld increase the reliability and validity. As explained, during the experiment the music was playing. Thus, when participants filled in the questinnaire they still culd hear the music. In research f Kappen (2014) a significant effect f music n prduct chice was fund when participants that did nt hear the music were excluded. In this experiment a relatively high amunt f peple heard the music and there was n difference when these peple were excluded. The fact that music was played while filling in the questinnaire might have had influence n answers in the questinnaire, while this was nt the aim. Fr example, it was asked whether peple heard music while ding grceries. It might be that peple than started listening and filled in yes while they did nt hear music when ding grceries. Filling in the questinnaire in anther rm is a slutin fr this limitatin. In the results became clear that 33% f the participants stated smething abut an r mre unrealistic aspect(s) f the virtual supermarket. Due t this striking number f 33 percent that nticed unrealistic aspects, it can be questined whether the results f this virtual supermarket are valid in a real supermarket envirnment. It might be that the virtual supermarket was t unrealistic and therefre behaviur changed. Research and tests can be dne in ther t find ut t which extent a real supermarket and virtual supermarket are cmparable. 5.3 Implicatins As described in the intrductin, there is nly little research dne yet cncerning the effect f strngly assciated music n prduct chice. This research can be cnsidered as a cntributin t a start pint fr further research in this field. In general, retailers cannt be advised t use musical assciatins in rder t influence the prduct chice f cnsumers. Hwever, it shuld be nted that this experiment nly investigated the effect f these music samples n a specific categry (meal packages). S, n cnclusins can be drawn n the effect n ther prduct categries. Because f that, retailers are advised t d mre research cncerning the effect f assciated music n prduct chice, befre a retailer might think f implementing this strategy. 45

47 References Agmn, E. (1990). Music thery as cgnitive science: sme cnceptual and methdlgical issues. Music Perceptin, Vlume 7(3), Alpert, I. J., Alpert, I.M. (1990). Music Influence n Md and Purchase intentins. Psychlgy and Marketing, Vlume 7(2), Alpert, I. J., Alpert, I.M., Maltz, E.N. (2005). Purchase ccasin influence n the rle f music in advertising. Jurnal f Business Research, Vlume 58, Areni, C. S., Kim, D. (1993). The Influence f Backgrund Music n Shpping Behavir - Classical Versus Tp-40 Music in a Wine Stre. Advances in Cnsumer Research, Vlume 20, Bacn, W.F., Egeth, H.E. (1994). Overriding stimulus-driven attentinal capture. Perceptin & psychphysics, Vlume 55(5), Bargh, J.A. (2006). What have we been priming all these years? On the develpment, mechanisms, and eclgy f nncnscius scial behavir. Eurpean Jurnal f Scial Psychlgy, Vlume 36, Batra, R., Ray, M.L. (1983). Advertising situatins: The implicatins f differential invlvement and accmpanying affect respnses. Infrmatin Prcessing Research in Advertising, Berger, J., Meredith, M., Wheeler, S. C. (2008). Cntextual priming: Where peple vte affects hw they vte. Prceedings f the Natinal Academy f Sciences f the United States f America, Vlume 105 (26), Billings, W.L. (1990). Effects f Stre Atmsphere n Shpping Behaviur. Hnrs Prjects. Paper Bitner, M.J. (1992). Servicescapes: the impact f physical surrundings n custmers and emplyees. The Jurnal f Marketing, Vlume 56, Cahill, L. McGaugh, J. L. (1998). Mechanisms f emtinal arusal and lasting declarative memry. TINS, Vlume 21 (7), Caldwell, C., Hibbert, S.A. (2002). The influence f music temp and musical preference n restaurant patrns behavir. Psychlgy f Marketing, Vlume 19, 895. Crawfrd, J.R., Henry, J. D. (2004). The Psitive and Negative Affect Schedule (PANAS): Cnstruct validity, measurement prperties and nrmative data in a large nn-clinical sample. British Jurnal f Clinical Psychlgy, Vlume 43, Danaher, P. J., Mullarkey, G.W. Factrs affecting nline advertising recall: A study f students. Jurnal f advertising research, Vlume 43 (3), Degenhardt, L. (2002). Why D Peple Act in Sustainable Ways? Results f an Empirical Survey f Lifestyle Pineers. Psychlgy f Sustainable Develpment,

48 Dnvan, R., Rssiter, J. (1982). Stre Atmsphere: An Envirnmental Psychlgy Apprach. Jurnal f Retailing, Vlume 58, Dwning, P. E. (2000). Interactin between visual wrking memry and selective attentin. Psychlgical Science, Vlume 11, Dubé, L., Mrin, S. (2001). Backgrund music pleasure and stre evaluatin Intensity effects and psychlgical mechanisms. Vlume 54, Duncan, J. (1984). Selective attentin and the rganizatin f visual infrmatin. Jurnal f Experimental Psychlgy, Vlume 113(4), Duncan, J., Humphreys, G. W. (1989). Visual search and stimulus similarity. Psychlgical Review, Vlume 96, Field, A. (2009). Discvering statistics using SPSS. Thusand Oaks: Sage Publicatins. French, A., Smith, G. (2013). Measuring brand assciatin strength: a cnsumer based brand equity apprach. Eurpean Jurnal f Marketing, Vlume 47 (8), Garlin, F. V., Owen, K. (2006). Setting the tne with the tune: A meta-analytic review f the effects f backgrund music in retail settings. Jurnal f Business Research, Vlume 59(6), Herringtn, J.D., Cappella, L.M. (1996). Effects f music in service envirnments: a field study. Jurnal f Services Marketing, Vlume 10 (2), Hsea, M. (2004). Let the music play. In-stre marketing, Jnides, J., Yantis, S. (1988). Uniqueness f abrupt visual nset in capturing attentin. Perceptin & Psychphysics, Vlume 43(4), Kappen, H. (2014). Nudging with Auditry Cues. Wageningen University. Keil, F.C. (1989). Cncepts, kinds, and cgnitive develpment. Cambridge: MIT Press. Ktler, P. (1973). Atmsphere as a Marketing Tl. Jurnal f Retailing, Vlume 49, Kunda, Z. (1999). Scial cgnitin: Making sense f peple. Cambridge: MIT Press. Lewis, A. (2008). The Cambridge handbk f psychlgy and ecnmic behaviur. Cambridge, Cambridge University Press: 1 nline resurce. Lftus, G.R. (1972) Eye fixatins and recgnitin memry fr pictures. Cgnitive Psychlgy, Vlume 3 (4), Lftus, F., Hffman, H., Lftus, G.R. (1991). Eye fixatins and memry fr emtinal events. Jurnal f Experimental Psychlgy, Vlume 17 (4), MacInnis, D.J., Park, C. (1991). The Differential Rle f Characteristics f Music n High- and Lw Invlvement Cnsumers Prcessing f Ads. Jurnal f Cnsumer Research, Vlume 18,

49 Martindale, C., Mre, K. (1988). Priming, prttypicality, and preference. Jurnal f Experimental Psychlgy: Human Perceptin and Perfrmance, Vlume 14, McCarrn, A., Tierney, K.J. (1989). The effect f auditry stimulatin n the cnsumptin f sft drinks. Appetite, Vlume 3 (2), McGldrick, P.J., Piers, C.P (1998). Atmspheres, pleasure and arusal: the influence f respnse mderatrs. Jurnal f Marketing Management, Vlume 14, Mehrabian, A. (1976) Public Spaces and Private Spaces: The Psychlgy f Wrk, Play and Living Envirnments, New Yrk: Basic Bks, Inc. Mehrabian, A., Russell, J. (1974). An Apprach t Envirnmental Psychlgy, Cambridge: MIT Press Mrris, R.K. Flk, J.R. (1998). Fcus as a cntextual priming mechanism in reading. Memry and Cgnitin, Vlume 26 (6), Myers, C.A. (2003). Managing brand equity: a lk at the impact f attributes. Jurnal f Prduct & Brand Management, Vlume 12(1), Niedenthal, P. M., Mndilln, L., Effrn, D.A., Barsalu, L.W. (2011). Representing Scial Cncepts Mdally and Amdally. In F. Strack, Förster J. (Ed.), Scial cgnitin The basis f human interactin. Nrth, A.C., Hargreaves, D.J., Mckendrick, J. (1999). The Influence f In-Stre Music n Wine Selectins. Jurnal f Applied Psychlgy, Vlume 84 (2), Oakes, S. (2000). The influence f the musicscape within service envirnments. Jurnal f Services Marketing, Vlume 14(7), Pieters, R., Warlp, L., Wedel, M. (2002). Breaking Thrugh the Clutter: Benefits f Advertisement Originality and Familiarity fr Brand Attentin and Memry. Management Science, Vlume 48 (6), Pieters, R., Wedel, M. (2004). Attentin capture and transfer in advertising: Brand, pictrial and textsize effects. Jurnal f Marketing, Vlume 68, Pieters, R., Wedel, M. (2006). Eye tracking fr Visual Marketing. In Fundatins and Trends in Marketing, Vlume 1 (4), Hanver: Nw Publishers. Psner, M. I. (1980). Orienting f attentin. Quarterly jurnal f experimental psychlgy, Vlume 32(1), Radcy, R.E., Byle, J.D. (1997). Psychlgical Fundatins f Musical Behavir. Springfield: Charles C. Thmas. Rethans, A. J., Swasy, J.L., Marks, L.J. Effects f televisin cmmercial repetitin, receiver knwledge, and cmmercial length: A test f the tw-factr mdel. Jurnal f Marketing Research, Rck, I., Gutman, D. (1981). The effect f inattentin n frm perceptin. Jurnal r Experimental Psychlgy: Human Perceptin and Perfrmance, Vlume 7(2),

50 Rumelhart, D.E., McClelland, J.E. (1986). Parallel distributed prcessing: Explratins in the micrstructure f cgnitin. Cambridge: MIT Press. Smith, E.E. (1990). Thinking: An invitatin t cgnitive science. Cambridge: MIT Press. Smith, P. C. & Curnw, R. (1966). 'Arusal hypthesis' and the effects f music n purchasing behavir. Jurnal f Applied Psychlgy, Vlume 50, Strebele, N., De Castr, J. M. (2004). Effect f ambience n fd intake and fd chice. Nutritin, Vlume 20(9), Sullivan M. (2002). The impact f pitch, vlume and temp n the atmspheric effects f music. Internatinal Jurnal f Retail & Distributin Management, Vlume 30 (6), Treisman, A.M., Gelade, G. (1980). A feature-integratin thery f attentin. Cgnitive psychlgy, Vlume 12(1), Vitz, P.C. (1972). Preference fr tnes as a functin f frequency (hertz) and intensity (decibels). Perceptin & Psychphysics, Vlume 11 (1), Watsn, D., Clark, L.A., Tellegen, A. (1988) Develpment and validatin f brief measures f psitive and negative affect: the PANAS scales. Jurnal f Persnality and Scial Psychlgy, Vlume 54 (6), Wilsn, S. (2003). The effect f music n perceived atmsphere and purchase intentins in a restaurant. Psychlgy f Music, Vlume 31, Wittgenstein, L. (1953). Philsphical investigatins. New Yrk: Macmillan. Wyatt, S.B., Winters, K.P., Dubbert, P.M. (2006). Overweight and Obesity: Prevalence, Cnsequences, and Causes f a Grwing Public Health Prblem. The American jurnal f the medical sciences, Vlume 331 (4), Yalch, R. F., Spangenberg, E. R. (2000). The effects f music in a retail setting n real and perceived shpping times. Jurnal f Business Research, Vlume 49(2), Yantis, S., Hillstrm, A.P. (1994). Stimulus-driven attentinal capture: evidence frm equiluminant visual bjects. Jurnal f Experimental Psychlgy: Human Perceptin and Perfrmance, Vlume 20(1), Yantis, S., Jnides, J. (1990). Abrupt visual nsets and selective attentin: Vluntary vs. autmatic allcatin. Jurnal f Experimental Psychlgy: Human Perceptin and Perfrmance, Vlume 16, Ynelinas, A.P. (2001). Cnsciusness, cntrl, and cnfidence: the 3 Cs f recgnitin memry. Jurnal f Experimental Psychlgy, Vlume 130 (3), Zaltman, G. (2003). Hw Custmers Think: Essential Insights int the Mind f the Market. Watertwn: Harvard Business Schl Press. 49

51 Appendixes Appendix 1: Sngs and YuTube links. Pre-test 1 Sng categry Vcal/Instrum ental Sng name YuTube link Mediterranean Instrumental Italian Medley ukljcwlapvg Mediterranean Vcal Teddy Ren - Picclissima Serenata 4mlSkPG2850 Mediterranean Vcal Claudi Villa Qui Stt Il Ciep Di Capri Rye0BNsibb0 Mediterranean Vcal Dmenic Mdugn Vlare vrf5zsczyes Mediterranean Instrumental Sunset in Capri (Tarzia/Ciplla) fjyb-fh8lw Asian Instrumental Dansngs Ambershire s-_1zltec8 Neutral Vcal Peter Nrth - Slw lve mfkl-bjaih8 Neutral Instrumental Backgrund Instrumental Music _s2yxraiwcq Pre-test 2 Neutral Instrumental Guitar backing track A minr/c Majr IFcZ_KZ5S4 Neutral Vcal Gerge Ezra - Listen t the Man ZS0WvzRVByg Neutral Instrumental Easy Listening Music: Guitar & Sax A9U5MDJ70I Neutral Instrumental Happy Acustic Backgrund Music - Shining Thrugh by 70VGIBu-nM Alum Neutral Instrumental Musse T - Oh Sng IxyOAtxsjvg 50

52 Appendix 2: Survey pre-test 1. The questins were asked fr all the eight sngs. 51

53 52

54 Appendix 3: Overview f bth pre-tests. T make all the results clear, a table with an verview f bth pre-test 1 as well as pre-test 2, is prvided belw. Sng 1: Italian Medley, sng 6: Dansngs Ambershire and sng 3: Easy Listening Music: Guitar & sax, were used in the virtual supermarket. Assciatin meals Assciatin cuntries Other factrs Pre-test 1 Mediterranean Asian American Others Mediterranean Asian American Others Musical Fit Arusal Knws Liking the sng sng 1 (Mediterranean) (20.57) (17.92) (25.83) sng 2 (Mediterranean) (20.99) (16.37) (22.54) sng 3 (Mediterranean) (18.51) (17.77) (22.9) sng 4 (Mediterranean) (22.99) (18.39) (21.97) sng 5(Mediterranean) (23.09) ( (21.41) sng 6 (Asian) (21.78) (17.75) (22.89) sng 7 (NA) (21.54) (20.80) (24.85) sng 8 (NA) (24.54) (15.31) (22.92) Pre-test 2 sng 1 (NA) (23.41) sng 2 (NA) (22.91) sng 3 (NA) (25.67) sng 4 (NA) (19.65) sng 5 (NA) (18.90) 30.21(12.29) (17.23) (18.12) (17.90) (14.73) (24.08) (12.01) (21.22) (12.04) (24.57) 53

55 Appendix 4: Survey pre-test vlume. The questins were asked fr all the three sngs. Luister naar nummer Wanneer vind u het vlume te zacht vr een supermarkt mgeving? Decibels: Cmputer: Wanneer vind u het vlume ged vr een supermarkt mgeving? Decibels: Cmputer: Wanneer vind u het vlume te hard vr een supermarkt mgeving? Decibels: Cmputer: 54

56 Appendix 5: Results pre-test vlume. Sng 1 Italian Medley Lw Middle High Mean: Sng 2 Dansngs - Ambershire Lw Middle High Mean: Sng 3 Easy Listening Music: Guitar & sax Lw Middle High Mean:

57 Appendix 6: Schedule. Gebuw (lkaal) Frum (0034) Frum (0034) Frum (0034) Frum (0034) Frum (0034) Datum Maandag 12 januari Dinsdag 13 januari Wensdag 14 januari Dnderdag 15 januari Vrijdag 16 januari Cnditie chtend Opbuwen Neutraal middel Aziatisch middel Aziatisch zacht Mediterraans middel Cnditie middag Neutraal zacht Aziatisch zacht Neutraal hard Neutraal hard Aziatisch middel Gebuw (lkaal) Frum (0034) Orin (G2010) Orin (G2010) Orin (G2010) Orin (G2010) Datum Maandag 19 januari Dinsdag 20 januari Wensdag 21 januari Dnderdag 22 januari Vrijdag 23 januari Cnditie chtend Aziatisch hard Opbuwen Neutraal zacht Aziatisch hard Neutraal hard Cnditie middag Aziatisch zacht Mediterraans zacht Mediterraans middel Neutraal middel Aziatisch middel Gebuw (lkaal) Orin (G2042) Orin (G2042) Orin (G2042) Orin (G2042) Orin (G2042) Datum Maandag 26 januari Dinsdag 27 januari Wensdag 28 januari Dnderdag 29 januari Vrijdag 30 januari Cnditie chtend Mediterraans zacht Mediterraans middel Neutraal middel Mediterraans hard Mediterraans hard Cnditie middag Neutraal zacht Aziatisch hard Mediterraans hard Mediterraans zacht 56

58 Appendix 7: Overview f prducts in the virtual supermarket. Asian prducts in the virtual supermarket: Cnimex F yng Hai Cnimex Babi Pangang Cnimex Babi Ketjap Cnimex Mihen Speciaal Cnimex Thaise Nasi Cnimex Tjap Tji Cnimex Kip Kerrie Madras Hnig Kip Tandri Hnig Nasi Speciaal Hnig Fijne Mie Mediterranean prducts in the virtual supermarket: Knrr Macarni Knrr Ristt Knrr Paella Knrr Spaghetti cn Funghi Knrr Ovenpasta Tmaat Mzarella Knrr Ovenpasta Carbnara Kaas Hnig Macarni venschtel Cabnara Hnig Tagliatella Rmsaus Hnig Lasagne Tradizinale Hnig Lasagnesaus Other prducts in the virtual supermarket: Maggi Hllands Pannetje Jachpannetje Sperziebnen Maggi Hllands Pannetje Champignns Prei Maggi Hllands Pannetje Rde Kl Appel Knrr Gegratineerde Witlf-Ham rlletjes Maggi Strganff Maggi Hllandse Aardappelgratin Kip Tmaat Maggi Zuurkl Craime Fraiche met gehakt uit de ven Maggi Prei Kerrie met gehakt uit de ven Maggi Witlf Spekjes met gehakt uit de ven Maggi Blemkl rm met gehakt uit de ven 57

59 Appendix 8: Cnsent frm. Infrmatiebrief nderzek - Virtuele supermarkt Waarm den we dit nderzek? Dit nderzek is bedeld m meer inzicht te krijgen in de keuzes die mensen maken bij het den van bdschappen in de supermarkt. Wie mag er meeden? Vrwaarden m me te kunnen den aan het nderzek zijn: Vldende kunnen hren Vldende kunnen zien Bereid zijn een testemmingsverklaring te tekenen Wanneer en waar is het nderzek? Deelname kan nu direct. Het nderzek neemt ngeveer 15 a 20 minuten in beslag. Het vindt plaats in lkaal 0034 p het universiteitsgebuw Frum f in lkaal G2010 p het universiteitsgebuw Orin. Wat hudt het nderzek in? Dit nderzek gaat ver het selecteren van prducten. Het bestaat uit de vlgende nderdelen: - Ondertekenen van een testemminsverklaring - Een krte intrductie - Het kiezen van een prduct in een virtuele mgeving (p de cmputer) - Een afsluitende vragenlijst Heel sms kan het bij het gebruik van de virtuele mgeving (p de cmputer) gebeuren dat mensen wat last krijgen van misselijkheid. Dit gebeurt vrnamelijk bij mensen die last hebben van wagenziekte en dan ng in zeldzame gevallen. Privacy Gegevens die verzameld zijn tijdens het nderzek zullen uiteraard vertruwelijk wrden behandeld. Vergeding Na aflp van het nderzek krijgt u een bedankje naar keuze, als dank vr uw deelname. Uiteraard kunt u p elk mment znder pgave van reden stppen met het nderzek. Dr wie wrdt dit nderzek uitgeverd? Dit nderzek wrdt uitgeverd dr: Wageningen UR, Marketing and Cnsumer Behaviur and Fd and Bibased Research: Jasper Freriks, jasper.freriks@wur.nl, Erica van Herpen, erica.vanherpen@wur.nl Anna Maaskant, anna.maaskant@wur.nl 58

60 Met eventuele vragen kunt u k bij hen terecht. Deelnemerslijst - Testemmingsverklaring Met de ndertekening van dit frmulier verklaart u deel te nemen aan het nderzek en geïnfrmeerd te zijn ver het del en de pzet van het nderzek. Naam Datum Tijd Handtekening 1. 59

61 Appendix 9: Instructins. Instructies navigatie: Het lpen dr de supermarkt kan dr middel van de spacenavigatr (3Dcnnexin) en de muis. De spacenavigatr werkt als een jystick en u kunt lpen naar de gewenste richting. Drmiddel van het indrukken van de spacenavigatr kunt u bukken en prducten die lager staan ged bekijken, als u de spacenavigatr mhg beweegt, gaat u weer naar de nrmale hgte. Met de muis kunt u klikken p een prduct en deze van dichterbij bekijken. U dient wel in de buurt van het prduct te staan m erp te kunnen klikken. Vervlgens kunt u besluiten (dr middel van klikken met de muis) dit prduct terug te leggen f te kpen. Als u een kruispunt nadert, kmen er pijlen in beeld. U kunt dan met de muis p de pijltjes klikken m een bepaalde kant p te gaan. Hierna kan je de Spacenavigatr weer gebruiken. Instructies bdschappen: Stelt u de vlgende situatie vr: u wilt een appel, een pak kekjes, een kuipje margarine (vr p brd) en als laatste een maaltijdmix kpen. U weet ng niet precies welke avndmaaltijd, maar u wilt een keer een maaltijd uit een pakje gebruiken. Kp in de virtuele supermarkt dus: 1 appel 1 pak kekjes 1 kuipje margarine (vr p brd) 1 maaltijdmix pakje naar keuze waarmee u wilt gaan kken. Ga er hierbij vanuit dat u de bijbehrende prducten later ng gaat verzamelen. Indien u alle prducten hebt geselecteerd, kunt u mij een seintje geven. U zult vervlgens een kleine vragenlijst invullen. Screensht schapverzicht virtuele supermarkt: In het nderstaande screensht is te zien van waar de prducten staan in de virtuele supermarkt. 60

62 Appendix 10: Survey virtual supermarket. Geachte meneer/mevruw, Bedankt vr het meeden aan dit nderzek! Nu vlgt een vragenlijst, deze zal ngeveer 10 minuten in beslag nemen. Uw annimiteit zal gewaarbrgd blijven. Alvast bedankt! Wat is uw respndentnummer? Welk maaltijdpakket heeft u gekzen? He velde u zich tijdens het rndlpen in de virtuele supermarkt? 61

63 He velde u zich tijdens het rndlpen in de virtuele supermarkt? In heverre hebben heeft u gelet p nderstaande aspecten tijdens het kiezen van het maaltijdpakket? (meerdere antwrden mgelijk) Heeft u ng p iets anders gelet? Nee Ja, namelijk In heverre heeft u p dit andere gelet? Zu u het gekzen prduct nrmaal k kpen? Zeer zeker niet Waarschijnlijk niet Neutraal Waarschijnlijk wel Zeer zeker wel Wat herinnert u zich van de supermarkt? Is u iets pgevallen aan de supermarkt? Schrijf hiernder alles p wat er in u pkmt. Heeft u muziek gehrd tijdens het uitkiezen van het prduct? Nee Ja Wat vr srt muziek heeft u gehrd? 62

64 He (n)prettig vnd u de muziek? He (n)aangenaam vnd u de muziek? Wat vnd u van het vlume van de muziek? In heverre bent u het eens met de vlgende stelling: "het vlume van de muziek vnd ik prima"? He passend (in een supermarkt mgeving) vnd u de muziek? He geschikt (vr in een supermarkt mgeving) vnd u de muziek? Was u bekend met het nummer? Nee Ja, ik heb het nummer wel eens gehrd, maar ken de titel en artiest niet Ja. ik ken de titel, maar niet de artiest Ja, ik ken de artiest, maar niet de titel Ja, ik ken de artiest en titel 63

65 Gebruikt u wel eens de maaltijdpakketten waar u in dit nderzek uit kn kiezen? En z ja, he vaak? Nee, ik gebruik deze maaltijdpakjes nit Ja, ik gebruik deze maaltijdpakjes minder dan 1 keer per jaar Ja, ik gebruik deze maaltijdpakjes 1 keer in de 6 maanden tt 1 keer per jaar Ja, ik gebruik deze maaltijdpakjes 1 keer in de 3 tt 6 maanden Ja, ik gebruik deze maaltijdpakjes 1 keer in de maand tt 1 keer in de 3 maanden Ja, ik gebruik deze maaltijdpakjes één keer in de twee weken tt 1 keer per maand Ja, ik gebruik deze maaltijdpakjes één keer per week tt 1 keer in de twee weken Ja, ik gebruik deze maaltijdpakjes 2 tt 5 keer per week Ja, ik gebruik deze maaltijdpakjes meer dan 5 keer per week Ja, ik gebruik deze maaltijdpakjes iedere dag U krijgt nu een aantal maaltijdpakketten te zien. U zult gevraagd wrden f u deze in de virtuele supermarkt hebt gezien f niet. Het laden van de plaatjes kan sms even duren. Heeft u nderstaand prduct gezien in de virtuele supermarkt? Ja Nee Heeft u nderstaand prduct gezien in de virtuele supermarkt? Ja Nee 64

66 Heeft u nderstaand prduct gezien in de virtuele supermarkt? Ja Nee Heeft u nderstaand prduct gezien in de virtuele supermarkt? Ja Nee Heeft u nderstaand prduct gezien in de virtuele supermarkt? Ja Nee 65

67 Heeft u nderstaand prduct gezien in de virtuele supermarkt? Ja Nee Heeft u nderstaand prduct gezien in de virtuele supermarkt? Ja Nee Heeft u nderstaand prduct gezien in de virtuele supermarkt? Ja Nee Heeft u nderstaand prduct gezien in de virtuele supermarkt? Ja Nee 66

68 Heeft u nderstaand prduct gezien in de virtuele supermarkt? Ja Nee Heeft u nderstaand prduct gezien in de virtuele supermarkt? Ja Nee Heeft u nderstaand prduct gezien in de virtuele supermarkt? Ja Nee Heeft u nderstaand prduct gezien in de virtuele supermarkt? Ja Nee 67

69 Heeft u nderstaand prduct gezien in de virtuele supermarkt? Ja Nee Heeft u nderstaand prduct gezien in de virtuele supermarkt? Ja Nee Heeft u nderstaand prduct gezien in de virtuele supermarkt? Ja Nee Heeft u nderstaand prduct gezien in de virtuele supermarkt? Ja Nee 68

70 Heeft u nderstaand prduct gezien in de virtuele supermarkt? Ja Nee Heeft u nderstaand prduct gezien in de virtuele supermarkt? Ja Nee Heeft u nderstaand prduct gezien in de virtuele supermarkt? Ja Nee Heeft u nderstaand prduct gezien in de virtuele supermarkt? Ja Nee Op de vlgende pagina's zullen ng enkele achtergrndvragen vlgen. 69