Effects of an evoked refreshing consumption context on hedonic responses to apple juice measured using best–worst scaling and the 9-pt hedonic category scale

Please download to get full document.

View again

of 5
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Effects of an evoked refreshing consumption context on hedonic responses to apple juice measured using best–worst scaling and the 9-pt hedonic category scale
  Short Communication Effects of an evoked refreshing consumption context on hedonicresponses to apple juice measured usingbest–worst scaling and the 9-pthedonic category scale Karen A. Lusk a, ⇑ , Nazimah Hamid b , Conor M. Delahunty c , Sara R. Jaeger d a Department of Food Science, University of Otago, PO Box 56, Dunedin 9054, New Zealand b Faculty of Health and Environmental Science, Auckland University of Technology, Private Bag 9006, Auckland 1142, New Zealand c Symrise Asia Pacific Pte Ltd, 226 Pandan Loop, Singapore 128412, Singapore d The New Zealand Institute for Plant & Food Research Ltd, Mt. Albert Research Centre, Private Bag 92169, Victoria Street West, Auckland 1142, New Zealand a r t i c l e i n f o  Article history: Received 3 December 2013Receivedinrevisedform18November2014Accepted 8 January 2015Available online 17 January 2015 Keywords: Scenario evoked contextConsumer hedonic testingAcceptance testingBest–worst scalingContext effects a b s t r a c t Awareness of the need to consider a product’s consumption context when measuring consumer hedonicresponse of a product is increasing among consumer sensory researchers. This study investigated theeffects of evoking a consumption context using a written scenario on hedonic response measured usingbest–worst scaling and the 9-pt hedonic category scale. Hedonic responses for four apple juices with rel-atively large sensory differences were compared when measured in the evoked context ‘when havingsomething refreshing to drink’ using best–worst hedonic scaling ( n  =65) and the 9-pt hedonic scale( n  =48). Best–worst scaling discriminated between the four apple juices when a refreshing contextwas evoked (  p  <0.01), while the juices were equally liked using the 9-point scale (  p  =0.41) when thesame context was evoked. Consumers perceived best–worst scaling to be more difficult than the 9-ptscale, however there was no difference between the two methods for consumers perceived accuracy of their liking information. The present study highlights that the effect of an evoked context on hedonicresponse may not be universal for hedonic methods. Further research is needed to understand the effectof evoking context on the liking of products, and to determine whether this measure reflects product lik-ing in an actual consumption context.   2015 Elsevier Ltd. All rights reserved. 1. Introduction Consumerhedonicresponsesarecommonlymeasuredinacon-trolled laboratory setting, despite the setting not being a typicalconsumption context for foods and beverages. Meiselman (2013)suggested that consumer sensory research will ‘move beyond thelaboratory’ and seek methodological developments that provide abetter balance between research carried out under controlled lab-oratory and natural settings. While visual, auditory, and physicalsurrounds have been used to create consumption contexts (e.g.Sester et al., 2013), physical modifications of the laboratory areexpensiveandtimeconsuming,andmaysurpriseaconsumermorethan evoking the intended consumption context (Petit &Sieffermann, 2007).Recently, written scenarios have been employed to evoke aproduct’s consumption context in the laboratory setting whenmeasuring consumer hedonics (e.g. Hein, Hamid, Jaeger, &Delahunty, 2010, 2012) and emotional responses (e.g. Piqueras-Fiszman & Jaeger, 2014). Unlike modification of the physical test-ingenvironmenttoevokeconsumptioncontexts, writtenscenariosallow consumers to individualize their consumption context byevoking their own consumption situation through imagery. Eachconsumer is able to imagine specific aspects of the consumptioncontext that are relevant to him or her and that define the focalcontext. Therefore, while the written scenario evokes a commoncontext, the details of that context are unique to the consumer.Evoking a context when a refreshing drink is desired, led to dif-ferences in hedonic ratings using the 9-pt hedonic scale for fourapple juices with subtle sensory difference compared to a controllaboratory setting (Hein et al., 2010). In another study, differences in liking for four apple juices with subtle sensory differences werefound when elicited in an evoked context of watching a movie atthe theatre, but not in the evoked context of eating breakfast ona weekend morning (Hein et al., 2012). Conversely, differences in liking for four blackcurrant juices with subtle sensory differences http://dx.doi.org/10.1016/j.foodqual.2015.01.0070950-3293/   2015 Elsevier Ltd. All rights reserved. ⇑ Corresponding author. E-mail address:  karen.lusk@otago.ac.nz (K.A. Lusk).Food Quality and Preference 43 (2015) 21–25 Contents lists available at ScienceDirect Food Quality and Preference journal homepage: www.elsevier.com/locate/foodqual  were not found when elicited in the evoked movie context, butwere established in the evoked breakfast context. The study alsofound that as perceived product appropriateness decreased in theevoked contexts, a general decrease in consumer liking wasobserved. Therefore product type, type of evoked consumptioncontexts, and the match between product and type of evoked con-text, are factors to be considered when hedonic responses are elic-ited using evoked consumption contexts.One factor that has not been explored in the application of evoked context is the use of different hedonic scaling methodolo-gies. Previous studies have only employed the 9-pt hedonic scale(Hein et al., 2010, 2012). Hein, Jaeger, Carr, and Delahunty (2008) demonstrated similar rank order for the liking of a selection of six commercially available muesli bars in a laboratory settingregardless of the scaling methods used (9-pt category, unstruc-tured line, labeled affective magnitude, preference ranking orbest–worst scaling). However, superior sample discriminationwas achievedwithbest–worst scaling and this has been supportedby other studies ( Jaeger & Cardello, 2009; Jaeger, Jørgensen,Aaslyng, & Bredie, 2008). In the previously mentioned studies byHein et al. (2010, 2012), not only were differences in hedonicresponses identify under an evoked context condition, consumersalso perceived the task of hedonic response as being easier thanwhen no context was evoked. While best–worst scaling better dis-criminatedbetween samples and was foundto be as easy to use asthe 9-pt scale when no context is evoked (Hein et al., 2008), the effect of evoking a consumption context when using hedonic scal-ing methods has yet to be explored.Against this background, the aim of the current study was toinvestigate the effect of evoking a consumption context using awritten scenario in the laboratory setting to elicit hedonic ratingsusing two different hedonic scaling methods: the 9-pt categoryand best–worst hedonic scaling methods. 2. Materials and methods  2.1. Participants A total of 113 consumers (18–65years old) evaluated a set of four apple juices in an evoked context ‘when having somethingrefreshing to drink’ using best–worst scaling ( n  =65, 62% female)or using the 9-pt category scale ( n  =48, 65% female). Ethicalapproval to perform the study was granted by the University of Otago, Human Ethics Committee. Participants were between 18and 65years of age, lived in New Zealand for 5 or more years, ingood health and indicated a liking for apple juice (Hein et al.,2012). Participants received a voucher to the local cinema for theirparticipation in the session.  2.2. Samples Four apple juices commercially available in the New Zealandmarket that had relatively large differences in perceived sensorycharacterwereselected.Thesesamples, referredtoasA1–A4, weredescribedbypanellistsinapilotstudyasbeingclearinappearancewith an artificial flavor and caramel aftertaste (A1); cloudy inappearance,lesssweetandmoresourthantheotherthreesamples(A2); clear in appearance with a honey odor (A3); and cloudy inappearance with a stewed apple flavor (A4). Samples A3 and A4were widely available in supermarkets and convenience storesaround New Zealand, while samples A1 and A2 were considered‘premium’ products and were available in fewer retail outlets.  2.3. Study design Using a between-subjects experimental design, consumerswere allocated to an evaluation session using either best–worstscaling or the 9-pt hedonic scale. All hedonic responses were mea-sured in the evoked context ‘when having something refreshing todrink’. The context was evoked using the same written scenario asused by Hein et al. (2012): ‘‘ Think about an occasion when you arehaving something refreshing to drink. Clearly imagine you are experi-encing this occasion. Now write a detailed description of the occasionthat you are imagining. ’’ Apple juice was selectedas the product forthis study as it represented an appropriate product-context match(Hein et al., 2012). The written scenario was projected on a screen and read aloud twice before consumers were requested to providea written response to the occasion that they were imagining. Onceconsumers provided a written response, instructions for productevaluation were given.Following a block design, consumers evaluating samples usingbest–worst hedonic scaling were presented with four triads of apple juice samples. Each of the four samples was presented toconsumers three times across the four triads. Within each triad,consumers were asked to identify which sample they liked themost and which they liked the least. Between triads, consumerswere asked to rinse their palate with water and take a self-admin-istered, one-min break. Sample presentations within and acrosstriads were balanced for presentation and carry over effects.Consumers using the 9-pt hedonic scale evaluated the samefour apple juice samples, once. The four apple juice samples wereserved and consumers were instructed to taste each sample andindicate their overall liking on 9-point hedonic category scaleslabeled from ‘like extremely’ to ‘disliked extremely,’ with all stan-dard incremental labels (Peryam & Pilgrim, 1957). They wereinstructed to take a one-min break between each sample and torinse their palate with water before assessing the next sample.Samples were coded with three digit numbers and presented in arandom order across consumers.Throughout evaluation of the samples, consumers wereinstructed to keep in mind the evoked context and were allowedto re-read their description of that occasion at any time. The writ-ten scenario was provided at the top of each page where hedonicratings were recorded. After evaluation of samples, consumerscompleted a questionnaire to explore whether the context waseffectively evoked, and how they felt about their evaluations(Hein et al., 2010, 2012). The questions were rated on 9-point cat- egory scales and read: (Q1) ‘ to what extent did the occasion youimagined compel you to desire something to drink? ’ (1=‘not at allcompelled,’ 9=‘very compelled’), (Q2) ‘ while you tasted the juicesamples, how vivid in your mind was the occasion you imagined? ’(1=‘not at all vivid,’ 9=‘very vivid’), (Q3) ‘ how easy/difficult did you find it to rate your liking/dislike of the juice samples? ’ (1=‘verydifficult,’ 9=‘very easy’), and (Q4) ‘ to what extent do you feel that the liking information you have given is accurate? ’ (1=‘not at allaccurate,’ 9=‘very accurate’).  2.4. Data analysis Best–worst data were tabulated for each judge by calculatingthe total number of times a sample was identified as being ‘mostpreferred’ and ‘least preferred.’ For each sample, and for each con-sumer, the number of times the sample was least preferred wassubtracted from the number of times it was most preferred. Thebest-minus-worst (B–W) scores were used as input data for analy-sis ( Jaeger & Cardello, 2009; Jaeger et al., 2008). The possible B–Wscores ranged from  3 to +3.Best–worst scaling and the 9-pt category scale were comparedfor their ability to detect hedonic difference between samples 22  K.A. Lusk et al./Food Quality and Preference 43 (2015) 21–25  when measured in the evoked context. This was achieved by ana-lyzing the data sets using analysis of variance. Sample wasincluded in the model for the best–worst data, as the nature of the data does not permit the inclusion of the participant effect.Sampleandparticipantwereincludedforthe9-pointhedonicdata.Tukey’s Honestly Significant Difference (HSD) test was carried outtodeterminesignificantdifferencesbetweensamplemeans.Signif-icance was tested at the 5% level.Consumers’ perception of how compelled they were to desire arefreshing drink (Q1) and context vividness (Q2) were obtainedfrom the questionnaire. Data collected using the best–worst and9-pt hedonic scaling methods were compared by analysis of vari-ance(methodasmaineffect)asamanipulationchecktoverifythatno difference between test groups existed with respect to how thecontext was evoked. Similarly, consumers’ perceptions of taskcompletion (Q3 & Q4) were compared between the two hedonicscaling methods. This was carried out to explore how the testmethods allowed consumers to articulate their liking/dislikingtoward samples. 3. Results and discussion  3.1. Evoked context manipulation check No difference was observed in how compelled consumers wereto desire a drink (Q1:  F  1,111  =1.35,  p  =0.25) or how vivid theirimagined occasions were (Q2:  F  1,111  =0.27,  p  =0.61) when usingthe best–worst and 9-pt category scales under the evoked refresh-ing context. Therefore, context was inferred to be similarly evokedfor the two hedonic methods and subsequent differences in hedo-nic responses could therefore be attributed to the scaling methodsused to elicit hedonic responses.  3.2. Hedonic ratings of apple juice samples A significant difference in hedonic response to the four apple juice samples was observed using best–worst scaling ( F  3,256  =8.25,  p  <0.01) under the evoked refreshing context (Fig. 1).Samples A3 and A4 were, on average, significantly more liked thanA1 and A2. There were no detected differences among the apple juice samples using the 9-pt category scale in the evokedrefreshing context condition ( F  3,141  =0.97,  p  =0.41, Fig. 2). Henceintheevokedcontextcondition,best–worstscalingbetterdetectedsample differences than the 9-pt category scale.Different conclusions were drawn regarding product likingwhenevokinga context, dependingonthe hedonicscalingmethodused. In this study, no difference in average liking of the apple juices was observed when measured under the evoked refreshingcontext using the 9-pt scale. However, the juices were, on average,liked differently when using best–worst scaling under the sameevoked context. Despite the differences in sample discrimination,the rank order of mean liking for the samples was similar for bothscalingmethods, withsamplesA1andA2lesslikedascomparedtosamples A3 and A4 (Figs. 1 and 2). Previous studies reported supe-rior sample discrimination using best–worst scaling when no con-text was evoked (Hein et al., 2008; Jaeger & Cardello, 2009). In this study, best–worst hedonic scaling demonstrated superior discrim-ination compared to the 9-pt hedonic scale when context wasevoked. However, this may be methodically related, as best–worstscale requires that each sample is tasted several times. Regardlessof context, samples are presented side-by-side with best–worstscaling, allowing consumers to directly compare the sensory prop-erties among samples, and make decisions about their likingaccordingly. The methodologyof best–worst scaling may thereforenot require context to further improvesample discrimination. Fur-ther research with the inclusion of a control condition for each of the hedonic scaling methods (i.e. no evoked context) would pro-videvalidationthatsimilarproductdiscriminationcanbeachievedregardless of whether or not context is evoked.Although improved discrimination by preference methods hasbeen demonstrated (relative to acceptance methods), this maydepend on the magnitude of sensory differences between samples.Villanueva,Petenate,anddaSilva(2005)reportedthatrankingwasbetter for sample sets with large sensory differences. However, ithas also been reported that preference ranking and the 9-pt cate-gory scale similarly discriminated orange juice samples with largesensory differences, while preference ranking better discriminatedapple juice samples with small sensory differences compared tothe 9-pt category scale (Barylko-Pikielna et al., 2004). As best– worst scaling is also used to measure preference and permitsside-by-side comparison of samples, sensitivity to sample differ-ence may depend on the magnitude of sensory differences among Fig. 1.  Mean B–Wscores and standard deviations for four commercial apple juices measured under an evoked refreshing context using best–worst scaling. Samples sharingthe same letter are not significantly different (  p  >0.05). K.A. Lusk et al./Food Quality and Preference 43 (2015) 21–25  23  samples. Previous studies comparing best–worst scaling to otherhedonic methods, including the present, have used samples withrelatively large sensory difference (Hein et al., 2008; Jaeger &Cardello, 2009). Jaeger and Cardello (2009) have suggested that further understanding on how best–worst scaling responds tosamples that are more similar, compared to different, is needed.Despite the relatively large differences in sensory propertiesamong the apple juices in the present study, consumers usingthe 9-pt hedonic scale, on average found them to be equallyacceptable in the evoked context. How well a product is liked isdependent on the context in which it is consumed (Edwards,Meiselman, Edwards, & Lesher, 2003; Meiselman, Johnson, Reeve,&Crouch,2000).Forexample,whiletwoproductsmaybelikeddif-ferently in the laboratory setting, the same two products may beequally liked in a focal consumption context. Such an effect mayalso be dependent on the type of consumption context that isevoked (Hein et al., 2012), where two products are equally liked in one context but not in another. Together, the present and previ-ous research, point to the existence of an interaction between thetype of consumption context, magnitude of product difference andtype of hedonic method used when measuring consumer hedonicresponse. Further research is needed to examine the impact of evoked context on hedonic methods when using a single productwith different sensory differences: one set having subtle sensorydifferences, andtheother large. It maybethecasethat whenseek-ing to understand consumer hedonic response for a product setwithsubtledifferenceinsensorycharacteristic, applyingahedonicscaling method known to improve sample discrimination hasmerit.  3.3. Consumers’ perception of task completion On average the 9-pt hedonic category scale (   x  =7.4±1.5) wasfound to be significantly easier for consumers to indicate their lik-ing/disliking toward samples than best–worst scaling(   x  =6.2±1.9) when in the evoked context (Q3:  F  1,111  =13.33,  p  <0.01). However, no difference in terms of consumers’ perceivedaccuracy of their liking ratings was found between the two scalingmethods (Q4:  F  1,111  =2.60,  p  =0.11).The fact that best–worst scaling was found to be more difficultforconsumerstousewhenindicatingliking/dislikingthanthe9-ptcategory scale (Q3) is contrary to previous studies which reportedbest–worst scaling to be equally easy to use as other scaling meth-ods (Hein et al., 2008; Jaeger & Cardello, 2009; Jaeger et al., 2008). Theperceivedincreaseddifficultyassociatedwithbest–worst scal-ing might be attributed to the application of an evoked contextcondition. Inthisway, not onlywereconsumersaskedtomakelik-ing/disliking decisions for three possible paired sample compari-sons within a triad, but consumers were also asked to take intoaccountcontextual informationduringtheirevaluations. However,the increased difficulty may also be related to fatigue due to thelarge number of samples and the repeated tasting of samplesrequired in best–worst scaling (Mueller, Francis, & Lockshin,2009). Future research exploring the influence of evoked contexton consumers perceived difficulty of best–worst scaling shouldinclude a comparison of liking scores that are elicited using best–worst scaling with and without an evoked context.The traditional laboratory test setting is a context where foodsand beverages are consumed however, that context is not one thatconsumers are accustom to eating in. Some researchers advocatemoving away from the traditional laboratory test setting and totest products in contexts intended for consumption (Meiselman,2013). Arguably, the advantage of evoked consumption context isthe fact that it is an approach that can be used to reduce the arti-ficialness of the laboratory test setting and provide an alternativecontext to consumers. When consumers enter the laboratory set-ting, each may bring with them a very different context or reasonfor the evaluation of samples. In some way, by providing consum-ers with a uniform frame of reference when they provide theirhedonic response, consumers may feel more confident in theirresponses, whereas in the absence of a consumption context, con-sumers may approach the hedonic task with a more analyticalmind-set.The present research demonstrates that the effect of an evokedcontext on hedonic response is not universal for hedonic methods.Whenmakingdecisionsregardingtheuseofanevokedcontextinacontrolled laboratory setting, one must carefully consider theobjective of the study, when deciding on the type of scalingmethod used and determining whether evoking a context is perti-nent to elicit consumer hedonic responses. An evoked context inthe laboratory setting may provide a better measure of the levelof consumer product liking when using the 9-pt hedonic scale.However, superior sample discrimination may be achieved withbest–worst scaling, regardless of whether measured using an Fig. 2.  Mean hedonic ratings and standard deviations for four commercial apple juices measured under an evoked refreshing context using the 9-pt hedonic category scale.Samples sharing the same letter are not significantly different (  p  >0.05).24  K.A. Lusk et al./Food Quality and Preference 43 (2015) 21–25  evoked context. Replication of the current findings is necessary tomore robustly understand the effect of evoked context on hedonicscaling method performance. Further research is also needed toinvestigate the effects on hedonic response when measured usingdifferent evoked contexts and other product categories varying indegree of sensory difference. References Barylko-Pikielna, N., Matuszewska, I., Jeruszka, M., Kozlowska, K., Brzozowska, A., &Roszkowski, W. (2004). Discriminability and appropriateness of categoryscaling versus ranking methods to study sensory preference in elderly.  FoodQuality and Preference, 15 , 167–175.Edwards, J. S. A., Meiselman, H. L., Edwards, A., & Lesher, L. (2003). The influence of eating location on the acceptability of identically prepared foods.  Food Qualityand Preference, 14 , 647–652.Hein, K. A., Hamid, N., Jaeger, S. R., & Delahunty, C. M. (2010). Application of awrittenscenariotoevokeaconsumptioncontextinalaboratorysetting:Effectson hedonic ratings.  Food Quality and Preference, 21 , 410–416.Hein, K. A., Hamid, N., Jaeger, S. R., & Delahunty, C. M. (2012). Effects of evokedconsumption contexts on hedonic ratings: A case study with two fruitbeverages.  Food Quality and Preference, 26  , 35–44.Hein, K. A., Jaeger, S. R., Carr, B. T., & Delahunty, C. M. (2008). Comparison of fivecommon acceptance and preference methods.  Food Quality and Preference, 19 ,651–661. Jaeger, S. R., &Cardello, A. V. (2009). Direct andindirect hedonic scaling methods: Acomparison of the labeled affective magnitude (LAM) scale and best–worstscaling.  Food Quality and Preference, 20 , 249–258. Jaeger, S. R., Jørgensen, A. S., Aaslyng, M. D., & Bredie, W. L. P. (2008). Best–worstscaling: An introduction and initial comparison with monadic rating forpreference elicitation with food products.  Food Quality and Preference, 19 ,579–588.Meiselman, H. L. (2013). The future in sensory/consumer research: Evolving to abetter science.  Food Quality and Preference, 27  , 208–214.Meiselman, H. L., Johnson, J. L., Reeve, W., & Crouch, J. E. (2000). Demonstrations of the influence of the eating environment on food acceptance.  Appetite, 35 ,231–237.Mueller, S., Francis, I. L., & Lockshin, L. (2009). Comparison of best–worst andhedonic scaling for the measurement of consumer wine preferences.  Australian Journal of Grape and Wine Research, 15 , 205–215.Peryam, D. R., & Pilgrim, F. J. (1957). Hedonic scale method of measuring foodpreferences.  Food Technology, 11 , 9–14.Petit, C., & Sieffermann, J. M. (2007). Testing consumer preference for iced-coffee:Doesthedrinkingenvironmenthaveanyinfluence? Food Quality and Preference,18 , 161–172.Piqueras-Fiszman, B., & Jaeger, S. R. (2014). The impact of evoked consumptioncontexts and appropriateness on emotion responses.  Food Quality andPreference, Part C, 32 , 277–288.Sester,C.,Deroy,O.,Sutan,A.,Galia,F.,Desmarchelier,J.F.,Valentin,D.,etal.(2013).‘‘Having a drink in a bar’’: An immersive approach to explore the effects of context on food choice.  Food Quality and Preference, 28 , 23–31.Villanueva, N. D. M., Petenate, A. J., &da Silva, M. A. A. P. (2005). Performance of thehybrid hedonic scale as compared to the traditional hedonic, self-adjusting andranking scales.  Food Quality and Preference, 16  , 691–703. K.A. Lusk et al./Food Quality and Preference 43 (2015) 21–25  25
Similar documents
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks