Effects of evoked consumption contexts on hedonic ratings: A case study with two fruit beverages

Food Quality and Preference 26 (2012) 35–44

Contents lists available at SciVerse ScienceDirect

Food Quality and Preference

journal homepage: www.elsevier .com/locate / foodqual

Effects of evoked consumption contexts on hedonic ratings: A case studywith two fruit beverages

Karen A. Hein a,⇑, Nazimah Hamid b, Sara R. Jaeger c, Conor M. Delahunty d

a Department of Food Science, University of Otago, P.O. Box 56, Dunedin 9054, New Zealandb School of Applied Sciences, Auckland University of Technology, Auckland 1010, New Zealandc The New Zealand Institute for Plant & Food Research Ltd., Private Bag 92169, Auckland Mail Centre, Auckland 1142, New Zealandd CSIRO Food and Nutritional Sciences, P.O. Box 52, North Ryde, NSW 1670, Australia

a r t i c l e i n f o a b s t r a c t

Article history:Received 12 January 2011Received in revised form 16 December 2011Accepted 23 February 2012Available online 3 March 2012

Keywords:ContextWritten scenarioConsumerHedonic rating

0950-3293/$ - see front matter � 2012 Elsevier Ltd. Adoi:10.1016/j.foodqual.2012.02.014

⇑ Corresponding author. Fax: +64 3 479 7567.E-mail address: [email protected]

Consumer liking is critical to the success of foods and beverages in the marketplace. In addition to thesensory properties, the context in which a product is consumed can influence how that product is liked.The aim of this paper was to examine the effect of different evoked consumption contexts on consumerhedonic responses. Written scenarios instructed consumers to imagine their own consumption context.The evoked consumption contexts were consumers imagined response to the written scenario. Meanhedonic ratings for four apple and four blackcurrant juices using three evoked consumption contextsand a control context (no evoked consumption context) were compared within and between the fourcontexts. The three evoked consumption contexts were having a fruit beverage: while eating breakfast,while watching a movie, and while having something refreshing to drink. The written descriptions pro-vided by the consumers established that different occasions were imagined for each of the three evokedconsumption contexts. A greater effect of evoked consumption contexts on hedonic response wasobserved for the blackcurrant juice compared to the apple juice. Lower hedonic ratings for the blackcur-rant juices occurred under the evoked contexts compared to the control context. Although differences inthe overall level of mean hedonic ratings and liking of individual blackcurrant juice samples wereobserved in evoked consumption contexts, rank order of the juice samples was similar. From the consum-ers’ perspective, evoking a context in a control setting did not impact on the task demands of hedonicratings. Findings suggest that hedonic ratings are influenced by the product type and the type of evokedconsumption context. This study demonstrated the potential use of evoked contexts in a laboratory set-ting as an approach to provide consumers with a frame of reference in which products are assessed. Evok-ing context using a written scenario may serve as a tool to explore product liking under differentconsumption contexts that may not be feasible to carry out due to practical or financial constraints.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

1.1. Importance of consumption contexts when measuring productliking

Bisogni et al. (2007) outlined eight different dimensions of eat-ing situations: food and drink, time, location, activities, social setting,mental processes, physical condition, and recurrence. Thereforethe consumption context of a product is not strictly based on thephysical setting or location, but is constructed of additional vari-ables related to the situation and occasion, the individual, and alsothe food (Meiselman, 1996). Consumer hedonic testing commonlytakes place in settings such as a controlled sensory laboratory, cen-tral location, or in the home. Testing in settings such as these have

ll rights reserved.

(K.A. Hein).

been criticised for not taking into account the foods’ consumptioncontext (Köster, 2003). For example, food is generally not con-sumed in an isolated booth as is commonly carried out in a con-trolled sensory laboratory. Foods and beverages are consumed indifferent consumption contexts and depending on those contexts,liking can change.

King, Weber, Meiselman, and Lv (2004) explored the effects ofsocial interaction, physical environment, food choice and the serv-ing of product as a meal component, on the liking of two variationsof salad, pizza and iced tea. Although these aspects of the con-sumption context were found to influence liking, the effects werenot consistent across food items. Literature has also demonstrateddifferences in hedonic ratings elicited in different eating situations,and situations with manipulated contextual variables (Bell &Meiselman, 1994; de Graaf et al., 2005; Edwards, Meiselman,Edwards, & Lesher, 2003; Hersleth, Mevik, Naes, & Guinard,2003; Hersleth, Ueland, Allain, & Naes, 2005; King, Meiselman,

36 K.A. Hein et al. / Food Quality and Preference 26 (2012) 35–44

Hollenstein, Work, & Cronk, 2007; Meiselman, Johnson, Reeve, &Crouch, 2000; Premavalli, Wadikar, & Nanjappa, 2009). Likingtherefore may best be measured in the product’s consumption con-text when contextual aspects beyond the physical location arepresent. The sensory laboratory is a context where foods and bev-erages are consumed. However, this context does not take into ac-count how and why the consumer might consume the test food. Asa product’s actual consumption context is not always possible orpractical to use for consumer testing, an approach is needed to ac-count for a product’s consumption context when carrying out he-donic testing in a laboratory test setting.

1.2. Method to evoke a product’s consumption context in thelaboratory setting

A written scenario has been used to evoke a consumption con-text in a laboratory setting (Hein, Hamid, Jaeger, & Delahunty,2010). The written scenario instructs a consumer to imagine an oc-casion when a food or beverage could be consumed and requiresthem to think of that occasion in their own mind. The evoked con-sumption context is the consumer’s response to the written sce-nario. Each consumer imagines their own personal occasion inresponse to the same written scenario given to all consumers. Inthe remainder of this paper, evoked consumption context will bereferred to as ‘evoked context’. The use of the evoked context, whenwanting something refreshing to drink, resulted in differences inapple juice hedonic ratings compared to the control laboratory set-ting when no consumption context was evoked (Hein et al., 2010).In addition, use of the written scenario to evoke the consumptioncontext did not impact upon the task related demands asked ofconsumers when providing their hedonic responses.

Other means of evoking contexts described the application ofauditory, visual and a combination of the two modes (Köster,2003). Although not discussed in detail, the paper indicated thatevoking context influenced product preferences. Visual, auditoryand olfactory cues have also been manipulated in the laboratorytest setting to evoke a hot context for the evaluation of iced coffeebeverages, although similar product liking was observed regardlessof the evoked context (Petit & Sieffermann, 2007).

The context in which a product is consumed can be specific tothe individual consumer. In addition, different products may beappropriate for the same consumption context. Therefore whenseeking to evoke a product’s consumption context in the laboratorysetting, it is important to have an approach that allows for differenttypes of consumption contexts to be evoked and allows consumersto personalise the context. Unlike manipulation by physical means(Petit & Sieffermann, 2007), written scenarios allow consumers topersonalise a product’s consumption context and may allow differ-ent consumption contexts to be evoked in the laboratory test set-ting (Hein et al., 2010).

1.3. Effect of consumption contexts on hedonic ratings

Differences in consumer hedonic responses for products havebeen observed dependent on consumption context. Boutrolle andDelarue (2009) have described ‘a change in a foods liking scoresfor the foods’, ‘a change in the degree of discrimination betweenthe foods’, and ‘a change in the hedonic ranking of the variousfoods compared’ as a result of eliciting hedonic response for foodsin different testing locations. Here these results are referred to aslevel, span and order effects. A level effect is observed when overallhedonic ratings increase (or decrease) with a context. For example,when three samples are rated 5, 6 and 7 in one context, and 7, 8and 9, respectively, in another. The second effect, span, occurs when

samples are liked in a similar order, but a difference in magnitudeis observed. For example, this is demonstrated if three samples arerated 5, 6 and 7 in one context, and 4, 6 and 8 in another. An indi-cator of a span effect is a change in discrimination for the differ-ences in sample liking. An order effect is observed when there isa product preference change. For example, sample A is liked morethan B in one context, while B is more liked than A in another con-text. An evoked context may uniquely impact hedonic ratings fordifferent products. While evoked contexts in a controlled labora-tory setting is a starting point to approximate a real consumptioncontext (Hein et al., 2010), further research is needed to under-stand whether different consumption contexts can be evoked usinga written scenario in the laboratory setting, and to examine howthese contexts influence hedonic responses.

1.4. Justification and aims of the study

The previous study by Hein et al. (2010) evoked one consump-tion context in the laboratory setting and demonstrated a changein hedonic response compared to a control context. In order toestablish the usefulness of evoked consumption contexts, it isimportant to demonstrate that consumption contexts of differenttype can be evoked using written scenarios, and to understandthe different effects of these contexts on hedonic ratings.

Hence the overall aim of this study was to examine the effect ofdifferent evoked consumption contexts on consumer hedonic re-sponses. To achieve this aim, three objectives were identified.The first objective was to determine whether different consump-tion contexts could be evoked in a laboratory setting using writtenscenarios. This would establish the efficacy of written scenarios toevoke different types of consumption contexts. The second objec-tive was to determine the effect of different evoked consumptioncontexts on hedonic ratings for two different fruit beverages. Thisobjective would provide an understanding of the effect of differentevoked consumption contexts on product liking, and how the effectvaried with product type. The third objective was to determinewhether evoked consumption contexts influenced how easy or dif-ficult consumers found it to evaluate their liking/disliking for prod-ucts. It may be the case that depending on the type of evokedconsumption context, hedonic ratings may differ in perceived dif-ficulty. Consumers may have a difficult task of rating products,and it is important to understand whether evoking a consumptioncontext influenced this task.

In this study, hedonic ratings for a set of apple and blackcurrantjuices were measured in a sensory laboratory in four different con-sumption contexts that included three evoked contexts: (1) break-fast on a weekend morning and having something to drink, (2)watching a movie at the theatre and having something to drink,and (3) when having something refreshing to drink; and (4) a con-trol setting where no context was evoked. We wanted to evokeconsumption contexts that are very different in the occasions thatconsumers bring to mind, to further explore the efficacy of writtenscenarios. Apple and blackcurrant juice samples were created tovary subtly in sensory differences. This was done to explorewhether differences in mean hedonic ratings observed among thethree evoked consumption contexts could be influenced by a dif-ference in the perception of underlying sensory character. If sam-ples with large differences were used (e.g. different flavours)instead, consumers may have developed preference for one prod-uct over another, and not be influenced by context. Also, productswith subtle differences would vary only in sensory character. Largedifferences could lead to a different type of product and change lik-ing for reasons other than sensory character. Apple juice was se-lected to replicate previous research that used the same samples

K.A. Hein et al. / Food Quality and Preference 26 (2012) 35–44 37

(Hein et al., 2010). Blackcurrant juice was used to further under-stand the effects of evoked contexts on different beverage types.

2. Materials and methods

2.1. Selection of written scenarios to evoke three consumption contexts

Three written scenarios that were developed to evoke contextsincluded eating breakfast on a weekend morning and having some-thing to drink, watching a movie at the theatre and having some-thing to drink, and when having something refreshing to drink.The present study sought to evoke consumption contexts varyingin different dimensions of eating (Bisogni et al., 2007). Breakfastis a consumption context defined as a meal and by the time ofday. Watching a movie in a theatre is a consumption context thatinvolves location and could be regarded as a social occasion,although not fixed in time across consumers. Consuming a refresh-ing beverage is a consumption context that pertains to satisfying aphysiological or psychological need. Although other evoked con-texts could have been studied, the interest of this study was notso much on the dimensions of eating situations of a context butthe fact that the contexts evoked were different.

Pilot work was carried out to verify that the consumption con-texts evoked by the three written scenarios would be: (i) appropri-ate when consuming apple juice and blackcurrant juice, (ii)different in the consumption contexts that they evoked, and (iii)easily interpretable by consumers. In the first pilot test, consumers(n = 17) were presented with each written scenario in a random or-der and were asked to imagine consuming a beverage in the con-sumption context specified, and to then provide a writtenresponse describing the occasion that they imagined. For eachevoked context, a questionnaire was administered to determinethat the beverages (apple and blackcurrant juices) were perceivedas being appropriate to consume in the occasions imagined. Forboth beverages, consumers responded to the question: ‘if providedto you, how appropriate would it be to consume apple juice/black-currant juice in the occasion you described’? As mean ratings wereabove five on the 9-point scale, the assumption was made thatboth beverages were appropriate to consume in the given contexts.Written responses provided by consumers indicated that the con-sumption contexts evoked were different for the three writtenscenarios. For the movie and breakfast contexts, consumerscommented that they were unsure if they should be imaginingan occasion when they want a beverage or are having a beverage.Subsequently, wording of all three written scenarios were revisedto evoke an occasion when consumers have a beverage. The follow-ing written scenarios were used in this study to evoke the break-fast, movie and refreshing contexts:

2.1.1. Breakfast context‘‘Think about an occasion when you are eating breakfast on a

weekend morning and having something to drink. Clearly imagineyou are experiencing this occasion. Now write a detailed descrip-tion of the occasion that you are imagining’’.

2.1.2. Movie context‘‘Think about an occasion when you are watching a movie at the

theatre and having something to drink. Clearly imagine you areexperiencing this occasion. Now write a detailed description ofthe occasion that you are imagining’’.

2.1.3. Refreshing context‘‘Think about an occasion when you are having something

refreshing to drink. Clearly imagine you are experiencing this

occasion. Now write a detailed description of the occasion thatyou are imagining’’.

2.2. Samples

Four samples each of two product categories were used: applejuice and blackcurrant juice. Within each product category, thefour samples were created to vary subtly in sensory character. Bysubtly varying the sensory characteristics of the juices it was as-sumed that the samples have similar liking in a controlled test set-ting. However depending on the evoked context, different sensorycharacteristics may be more or less liked. Four samples were usedfor each fruit beverage as this was an acceptable number for con-sumer testing with respect to sensory and mental fatigue.

2.2.1. Apple juiceApple juice samples were those used in the previous study by

Hein et al. (2010). Using a base apple juice (Fresh-Up Crisp AppleJuice, Frucor Beverages Ltd., Auckland, New Zealand), four sampleswere created using a 2 � 2 design consisting of citric acid andstrawberry flavour (Hansells Natural Strawberry Flavour, Old Fash-ioned Foods Ltd., Auckland, New Zealand). These samples were re-ferred to as A (0.12 g citric acid/100 ml, 0 ml strawberry flavour), B(0.12 g citric acid/100 ml, 0.05 ml strawberry flavour/100 ml), C(0 g citric acid, 0 ml strawberry flavour) and D (0 g citric acid,0.05 ml strawberry flavour/100 ml). Both apple and blackcurrantjuices were prepared the day before evaluation and stored refriger-ated (�4 �C). Prior to evaluation, 30 ml of samples were pouredinto plastic portion cups with lids. Small amount of samples wereused, as is typical in central location tests. Samples were placed inan insulated container with crushed ice, and stored refrigerateduntil further evaluation. Samples were coded with three digit ran-dom numbers and the order of presentation was balanced (MacFie,Bratchell, Greenhoff, & Vallis, 1989).

2.2.2. Blackcurrant drinkFour blackcurrant juice samples were created using a blackcur-

rant concentrate (Barker’s, Unsweetened Blackcurrant Juice, Geral-dine, NZ). A 2 � 2 design was applied to produce samples varyingin added sucrose and orange flavour (Hansells Natural Orange Fla-vour, Old Fashioned Foods Ltd., Auckland, New Zealand). Thesesamples will be referred to as E (12.5 g concentrate, 12.3 g sucrose,0 ml orange flavour/100 ml water), F (12.5 g concentrate, 12.3 g su-crose, 0.01 ml orange flavour/100 ml water), G (12.5 g concentrate,10 g sucrose, 0 ml orange flavour/100 ml water), and H (12.5 g con-centrate, 10 g sucrose, 0.01 ml orange flavour/100 ml water). Sam-ples G and H were prepared to a final solution of �12�Brix. Thislevel was used as the level of sweetness for the base juice, giventhat a selection of blackcurrant juices available in New Zealanddemonstrated a range of between 8� and 13� Brix. Samples werepresented as described in Section 2.2.1.

Pilot testing was performed to determine whether the level ofadded sucrose and orange flavour made samples just noticeablydifferent in sensory character. Two separate ascending forcedchoice tests (3-AFC) were conducted with different consumersfor determination of added levels of sucrose (n = 16) and orangeflavour (n = 15) (Lawless & Heymann, 1999). The group thresholdfor sucrose and for orange flavour was calculated by taking thegeometric mean of the individual consumer thresholds (Lawless& Heymann, 1999). An additional 2.3 g/L of sucrose in the basejuice from 10 g/L, to make 12.3 g/100 ml, and 0.008 ml/100 ml or-ange flavouring in the base juice were determined. A slightly high-er value of 0.01 ml/100 ml was used for orange flavour in the finalmanipulation of samples. This was to ensure that all consumerswould perceive slight differences.

38 K.A. Hein et al. / Food Quality and Preference 26 (2012) 35–44

2.3. Main study

2.3.1. ConsumersConsumers (n = 257: 66% female, 34% male) were between 18

and 65 years of age, lived in New Zealand for five or more years,were in good health and liked both apple and blackcurrant juices.For consumers to be considered as liking the beverages, they wererequired to provide a liking response of five (neither like nor dis-like) or greater on the 9-point hedonic scale for both apple andblackcurrant juices without tasting involved. This was done toscreen against those who disliked apple and blackcurrant juices.Consumers meeting the criteria were invited to participate in thestudy. Consumers were required to attend one evaluation session,carried out in individual booths. All sessions were held at 12 pmfrom June to September, 2009. Ethical approval to perform thestudy was granted by the University of Otago Human EthicsCommittee.

A between subjects design was applied so that consumer groupswere exposed to one of the four consumption contexts. Thereforeall consumers would have been naïve in terms of the consumptioncontexts, sample sets and study design. Consumers were assignedto the control context (n = 64), evoked breakfast context (n = 63),evoked movie context (n = 68) and evoked refreshing context(n = 62). For each consumption context the consumer groups werebalanced with respect to age (v2

12 = 3.750, p = 0.988) and gender(v2

3 = 0.972, p = 0.808). By demonstrating homogeneity in keydemographics, we infer that subsequent differences in resultsfound between contexts arise from the experimental consumptioncontexts.

2.3.2. Control context (no evoked context)Consumers were asked to attend a one-hour evaluation session.

Four apple juice samples were served and consumers were in-structed to taste each sample and indicate their overall liking on9-point hedonic category scales labelled from ‘like extremely’ to‘disliked extremely’, with all standard incremental labels (Peryam& Pilgrim, 1957). They were instructed to take a one-minute breakbetween each sample and to rinse their palate with water beforecontinuing.

When all consumers had completed their evaluations, sampletrays and evaluation forms were collected. Blackcurrant juice sam-ples were then distributed. These samples were evaluated usingthe same procedure as the apple juice. Apple juice was always pre-sented first to consumers in the three evoked contexts and the con-trol context. This was done to replicate, and enable comparisonwith previous research that used the same apple juice samples(Hein et al., 2010). Apple juice was presented first in order to notexperience order effects. After evaluation of both beverages, con-sumers completed a questionnaire to indicate how they felt abouttheir evaluations. Consumers responded to the questions, ‘howeasy/difficult did you find it to rate your liking/dislike of the juicesamples’? (1 = ‘very difficult’, 9 = ‘very easy’), and ‘to what extentdo you feel that the liking information you have given is accurate’?(1 = ‘not at all accurate’, 9 = ‘very accurate’). It is important toacknowledge that the sensory laboratory is a context for the eval-uation of products. However the sensory laboratory is not a contextwhere consumers naturally eat and drink products. Like all physi-cal locations, the sensory laboratory is a context where consumersmay have expectations. In the control context, no consumptioncontext was evoked using written scenarios.

2.3.3. Evoked consumption contextsThe same sample evaluation procedure as described for the con-

trol context was followed with the following exceptions. Beforesample evaluation instructions were given, the written scenariowas presented (Section 2.1). The written scenario was projected

on a screen and read aloud twice before consumers were requestedto provide a written response to the occasion that they were imag-ining. Once all consumers had completed their response, instruc-tions for product evaluation were given. Consumers wereinstructed to keep in mind the occasion that they had describedthroughout evaluation of the samples and were allowed to re-readtheir description at any time. The written scenario was provided atthe top of each page where hedonic ratings were recorded.

Consumers completed a questionnaire following product evalu-ation. In addition to the two task demand questions asked in thecontrol context, three additional questions were included on thequestionnaire to explore whether context was effectively evoked.The questions were rated on 9-point category scales and read: ‘towhat extent did the occasion you imagined compel you to desire some-thing to drink’? (1 = ‘not at all compelled, 9 = ‘very compelled’),‘while you tasted the juice samples, how vivid in your mind was theoccasion you imagined’? (1 = ‘not at all vivid’, 9 = ‘very vivid’), and‘if provided to you, how appropriate would it be to have apple juice/blackcurrant juice in the occasion that you imagined’? (1 = ‘not atall appropriate’, 9 = ‘very appropriate’). The first two questionsserved to understand how well the context was evoked from theconsumer’s perspective. The last question served to understandwhether the beverages would be perceived as appropriate to con-sume in the participant’s imagined occasion.

2.4. Data analysis

2.4.1. Effectiveness of evoking different consumption contexts usingwritten scenarios

Representative examples of written descriptions provided byconsumers in each of the three evoked consumption contexts werecollated. This was done to achieve the first objective by ascertain-ing whether different consumption contexts were effectivelyevoked for the three written scenarios.

To further achieve the first objective, questionnaire data regard-ing consumers’ perceptions of how the contexts were evoked wasanalysed, exploring whether context was effectively evoked amongthe three written scenarios. A one-way analysis of variance wasperformed to compare responses among the three evoked contextsfor each of the four questions asked on the questionnaire. Tukey’spost hoc testing was carried out where significant differences werefound (p < 0.05).

2.4.2. Effectiveness of different evoked consumption context onhedonic ratings

To achieve the second objective, hedonic ratings for the appleand the blackcurrant juices elicited in the three different evokedcontexts and control context were compared within each juicetype in terms of (i) mean hedonic ratings, (ii) ability to detect dif-ferences in product liking and (iii) product preferences.

Mean hedonic ratings among the four consumption contextswere compared by one-way analysis of variance for each beveragetype. This was to determine whether different levels of overallproduct liking resulted depending on the consumption context.

The ability to detect sample difference (p < 0.05) was tested bytwo-way analysis of variance (sample and consumer effects) with-out interaction with hedonic rating data from each consumptioncontext for each beverage type. Tukey’s post hoc testing was car-ried out on significant sample effects to identify samples that werestatistically different from one another.

Paired samples t-tests were performed to further compare dif-ferences in sample liking for the six possible sample pairs withineach consumption context for each beverage type. Resulting t-val-ues were compared among the contexts, as the relative size of t isan indicator of the magnitude of sample discrimination. A largervalue of t observed in one context compared to another, would

K.A. Hein et al. / Food Quality and Preference 26 (2012) 35–44 39

suggest a greater magnitude of discrimination between contexts.Authors acknowledge that different consumer numbers were usedin the different consumption contexts, with the greatest numberused in the evoked movie (n = 68) and the least used in the evokedrefreshing context (n = 62). However, as the difference in consumernumbers was negligible, there would be little or no impact onvalidity of comparisons made.

Separately for each beverage type, a two-way analysis of vari-ance with context and juice samples as main effects was carriedout, whilst testing for a significant interaction. This was done to ex-plore whether the order of sample liking changed among the fourconsumption contexts. An interaction (p < 0.05) would indicatethat samples were liked differently depending on the consumptioncontext.

2.4.3. Consumers’ perceptionsThe third objective of this research was achieved by exploring

consumers’ perceptions of perceived difficulty of the hedonic taskand accuracy of hedonic information obtained from the question-naire for each of the four consumption contexts. Separate one-way analyses of variance were performed to compare responsesfor these two questions between the four consumption contexts.Tukey’s post hoc testing was carried out where significant differ-ences were found (p < 0.05). Consumer’s perception provided anunderstanding as to whether evoking a consumption context influ-enced the hedonic task.

3. Results

3.1. Effectiveness of evoking different consumption contexts usingwritten scenarios

Representative verbatim examples of written description ofoccasions imagined by consumers for each of the evoked contextsare presented in Table 1. The purpose of exploring the writtendescriptions provided by consumers was to ascertain that the con-sumer’s imagined occasion was representative of the intended con-sumption context. The written descriptions indicated thatconsumers imagined and described occasions for the respectivecontexts adequately. In the case of the movie context however,58% of the consumers described consuming a carbonated beverage(e.g. coke, diet coke, fizzy drink, fanta, etc.).

Questionnaire data indicated that consumers in the movie con-text were significantly less compelled to desire something to drinkcompared to those in the breakfast and refreshing contexts(F2,190 = 11.75, p < 0.001). Consumers in the refreshing context feltthat the imagined occasion was more vivid when evaluating thesamples (F2,190 = 5.47, p = 0.005), compared to consumers in themovie context (Table 2). Both apple (F2,190 = 22.43, p < 0.001) andblackcurrant (F2,190 = 9.15, p < 0.001) juices were significantly lessappropriate to consume in the movie context than in the refreshingor breakfast contexts (Table 2). These results indicated differencesexisted in how the three consumption contexts were evoked, andthe level of appropriateness to consume the juice products in eachcontext.

3.2. Effectiveness of different evoked consumption contexts on hedonicratings of apple juice

3.2.1. Mean hedonic ratingsNo significant difference in the overall level of mean hedonic

ratings for apple juice was observed among the four consumptioncontexts (F3,1024 = 1.48, p = 0.219). Results indicated that applejuice was liked similarly regardless of whether context was evoked.

3.2.2. Ability to detect differences in product likingAnalysis of variance showed no significant difference in mean

hedonic ratings among the apple juice samples in the control(F3,189 = 1.25, p = 0.294), refreshing (F3,183 = 0.25, p = 0.860), andthe breakfast contexts (F3,186 = 0.64, p = 0.592). However a signifi-cant difference among samples was observed in the movie context(F3,201 = 2.65, p = 0.050). In the movie context, the sample contain-ing added citric acid and strawberry flavouring (B) was the leastliked and was significantly different from the sample containingadded strawberry flavour (D), the most liked sample (Fig. 1).

In the movie context, the paired-samples t-test discriminatedthree of the possible six sample pairs, with the sample containingadded strawberry flavouring (D) being more liked than the samplescontaining added citric acid (A, p = 0.016), both added citric acidand strawberry flavouring (B, p = 0.009), and without anythingadded (C, p = 0.057) (Fig. 2). No pair-wise sample discriminationwas observed in the refreshing or breakfast contexts (p > 0.10).However, the sample with added strawberry flavouring (D) wassomewhat more liked (p = 0.077) than the samples with bothadded citric acid and strawberry flavour (B) in the control context.Overall, the apple juices was better differentiated for liking in themovie context compared to the refreshing context, breakfast con-text and control context.

3.2.3. Order of sample likingNo significant interaction was observed between the four apple

juice samples and the four consumption contexts (F9,1012 = 0.39,p = 0.940). This indicated that the four apple juices were liked ina similar rank order among the four contexts. Therefore, whilesamples were liked differently in the movie context (Section3.2.2), no overall statistical difference was evident in terms ofhow samples were liked among the three evoked contexts and con-trol. However, noticeably greater hedonic ratings resulted for sam-ples A, B and C in the refreshing context compared to the controlcontext (Fig. 1). Sample B demonstrated the greatest increase andwas rated 0.4 point higher on the 9-point hedonic scale in therefreshing context compared to the control context.

3.3. Effectiveness of different evoked consumption contexts on hedonicratings of blackcurrant juice

3.3.1. Mean hedonic ratingsA significant difference in the level of mean hedonic ratings for

blackcurrant juice was observed among the four consumption con-texts (F3,1024 = 7.51, p < 0.001). Overall, blackcurrant juice was mostliked in the control context, followed by the breakfast context, therefreshing context, and was least liked in the movie context. Spe-cifically, blackcurrant juice was significantly less liked in the movie(�x ¼ 5:8) than the breakfast (�x ¼ 6:3) and control (�x ¼ 6:5) con-texts. Blackcurrant juice was significantly less liked in the refresh-ing context (�x ¼ 6:0) than the control context. Blackcurrant juicewas less liked when evaluated in the evoked contexts comparedto the control context.

3.3.2. Ability to detect differences in product likingAnalysis of variance determined that the four blackcurrant

juices were equally liked in the movie (F3,201 = 0.643, p = 0.59)and control (F3,189 = 2.31, p = 0.078) contexts, while they were likeddifferently in the breakfast (F3,186 = 4.31, p = 0.006) and refreshing(F3,183 = 3.40, p = 0.019) contexts. In the breakfast context, the sam-ple without added sugar or orange flavouring (G) was significantlymore liked than the sample with added orange flavouring (H)(p < 0.05). In the refreshing context, the sample with added sugar(E) was significantly more liked than the sample with added or-ange flavouring (H) (Fig. 3).

Table 1Verbatim examples of consumer responses to the written scenarios used to evoke the contextual occasions: while eating breakfast, while watching a movie, and while havingsomething refreshing to drink.

Evoked contexts

Breakfast Movie Refreshing

Having breakfast on a sunny balcony. To include jam,marmalade rolls, coffee, fruit juice and water. Thebalcony has a sea view and I am reading thenewspaper. No traffic or other noise, nice andpeaceful. No other people or distractions. No radio ormusic

Sitting in the seats watching an action movie with myfriends most likely William. We are sharing a largecoke with no ice. We are sitting in the middle near theback and are stretched out comfortable. It is dark andwarm and the movie is very enjoyable

Drinking from a water bottle or a bottle of some coldbeverage mid-way up walking the Bethune’s Gullywalking track (Dunedin). Cold wind but overly hotfrom the hike and stopping to have a drink. Not anathletic hike but more of a recreational walk up a hill.Probably standing up and a bit exhausted

Breakfast in lounge, sitting on couch, watching TV.Eating white toast which has been just ‘browned’with melting margarine on top. With 2 poached eggs,that was cooked in boiling water, with a dash(teaspoon) of vinegar added to it. Salt on top of theeggs on toast. With a cup of herbal tea. Rooibos (southAfrican) tea) with one sugar and milk.

I’m at the movies, it’s about half way throughscreening. Now the particular movie is quite boring.Unsurprising I’m getting restless. I look through thefood and drink I have, I always seem to do this whenI’m bored and reach for the lemonade and the packetof fruit burst. I’ll probably go through this at intervalsto about 10 min before the movie ends.

After working outside cutting firewood for 3 h on anovercast day. While not a hot day I was tired andexhausted and enjoyed a cold bottle of PowerAid atthe end of the job.

I am sitting in a small café overlooking the ocean,reading the Saturday paper eating a smoked salmonbagel, drinking coffee and orange juice. I am with mypartner. There is soothing music being played and theocean is calm. It is very sunny

I am at the movie theatre. I am probably drinking alarge lift, no ice. The movie is probably some sort ofaction film although if I went with my friends itmaybe a chick flick/period drama. It would mostlikely be on a Sunday afternoon and I would possiblybe hung-over. I would take my shoes off and curl up inthe seat. I like the smell of popcorn but would nothave bought any

It is swelteringly hot summers day, and there is nobreeze. I have a cold drink and can hear the ice cubesclinking in the glass and feel very refreshed

I would be at home with my mum and dad, probablyaround 11 am. I would usually have a cookedbreakfast and my mum would make me a cup ofEnglish breakfast tea with a small amount of trimmilk and no sugar

Going out with my boyfriend, watching a chick flick.We will be seated about 2/3 of the way back from thescreen. We will be drinking frozen coke and eatingpopcorn. Feeling quite relaxed and happy, lookingforward to enjoying a brainless movie

It is a hot summer’s day and I am sitting in my deckenjoying an ice cold beer with my brother. The sun isshining and it is 25 �C, not a breath of wind in the air.We relax in our deck chairs as we sip our beer

Table 2Mean ratings and standard deviation from the consumer questionnaire for each consumption context. Contexts were compared by analysis for variances and Tukey’s post hoc testperformed for significant questions (p < 0.05).

Consumer perceptions Test conditions (mean (standard deviation))

Control+(n=64) Refreshing (n = 62) Breakfast (n = 63) Movie (n = 68)

Easy/difficult to rate liking/disliking. (F3,253 = 2.029, p = 0.110) 6.1 (2.2) 7.0 (2.0) 6.8 (1.9) 6.8 (2.2)Accuracy of liking information. (F3,252 = 0.526, p = 0.665)^ 7.4 (1.2) 7.6 (1.3) 7.5 (1.3)* 7.5 (1.2)Compelled to desire a drink. (F2,190 = 11.745, p < 0.001) – 7.6a (1.4) 6.8a (1.8) 5.9b (2.4)Vividness of imagined occasion. (F2,190 = 5.465, p = 0.005) – 7.0a (1.2) 6.3ab (1.8) 6.0b (2.1)Appropriateness of apple juice inimagined occasion. (F2,190 = 22.433, p < 0.001) – 7.7a (1.8) 6.8a (2.5) 5.0b (2.6)Appropriateness of blackcurrant juice inimagined occasion. (F2,190 = 9.147, p < 0.001) – 5.8a (2.3) 5.5a (2.6) 4.1b (2.4)

* n = 62 as one consumer did not respond to question.+ Test conditions with different letters within a row are significantly different (p < 0.05).^ One consumer failed to respond to questionnaire.

5

7

9

Mea

n he

doni

c ra

ting

b

a

1

3

A B C D

Breakfast (n=63)Control (n=64)Movie (n=68)*Refreshing (n=62)

*Letter indicate significantly different samples ( p<0.05 ).

Fig. 1. Mean hedonic ratings elicited in the four consumption contexts for fourapple juices (1 = dislike extremely, 9 = like extremely). Sample A = 0.12 g citric acid,0 ml strawberry flavour/100 ml juice, sample B = 0.12 g citric acid, 0.05 ml straw-berry flavour/100 ml juice, sample C = 0 g citric acid, 0 ml strawberry flavour/100 ml juice, and sample D = 0 g citric acid, 0.05 ml strawberry flavour/100 ml juice.

40 K.A. Hein et al. / Food Quality and Preference 26 (2012) 35–44

In the breakfast context, the paired samples t-test discriminatedthree pairs of samples with respect to liking (Fig. 4). Specifically,the sample without added sugar or flavouring (G) was more likedthan the sample with added orange flavouring (H, p < 0.001), andmore than the samples with added sugar (E, p = 0.072). Also inthe breakfast context, the sample containing both added sugarand orange flavouring (F) was more liked than the sample withonly orange flavouring (H, p = 0.016). In the refreshing context,pair-wise sample discrimination of the blackcurrant juices wasalso observed for three pairs of samples, with sample E more likedthan sample G (p = 0.095) and H (p = 0.003), and sample F moreliked than H (p = 0.043) (Fig. 4). In the control context, samples F(p = 0.084) and E (p = 0.035) were more liked than H. For all con-sumption contexts, samples G and F, and samples E and F werenot discriminated.

For the movie context, the sample with added orange flavouring(H) was liked over samples E (p = 0.433), F (p = 0.241) and G(p = 0.211), although not significantly (Fig. 4). The trend in theother three consumption contexts was opposite, where sample Hwas the least liked compared with sample E, F and G.

Refreshing

B

C

D

C

A

A

A

B

Preferred Sample Preferred Sample

-4 -2 0 2 4t-value

Movie

Control

BreakfastD

D

B

C

p=0.05

Fig. 2. t-Values comparing hedonic ratings for all possible pairs of four apple juicesevaluated using three evoked contexts (refreshing, movie, breakfast) and withoutan evoked context in a controlled laboratory setting. Sample A = 0.12 g citric acid,0 ml strawberry flavour/100 ml juice, sample B = 0.12 g citric acid, 0.05 ml straw-berry flavour/100 ml juice, sample C = 0 g citric acid, 0 ml strawberry flavour/100 ml juice, and sample D = 0 g citric acid, 0.05 ml strawberry flavour/100 ml juice.

3

5

7

9

Mea

n he

doni

c ra

ting

Breakfast (n=63)*

b

A

Ba

1

3

E F G H

Control (n=64)Movie (n=68)Refreshing (n=62)

Capital and lower case letters indicate significantly different samples, respectively ( p<0.05 ).

Fig. 3. Mean hedonic ratings elicited in the four consumption contexts fourblackcurrant juices (1 = dislike extremely, 9 = like extremely). Sample E = 12.5 gconcentrate, 12.3 g sucrose, 0 ml orange flavouring/100 ml water, sample F = 12.5 gconcentrate, 12.3 g sucrose, 0.01 ml orange flavouring/100 ml water, sampleG = 12.5 g concentrate, 10 g sucrose, 0 ml orange flavouring/100 ml water, andsample H = 12.5 g concentrate, 10 g sucrose, 0.01 ml orange flavouring/100 mlwater.

Refreshing

H

H

F

E

G

F

G

H

Preferred Sample Preferred Sample

-4 -2 0 2 4t-value

Movie

Control

Breakfastp=0.05

E

E

G

F

Fig. 4. t-Values comparing hedonic ratings for all possible pairs of four blackcurrantjuices evaluated using three evoked contexts (refreshing, movie, breakfast) andwithout an evoked context in a controlled laboratory setting. Sample E = 12.5 gconcentrate, 12.3 g sucrose, 0 ml orange flavouring/100 ml water, sample F = 12.5 gconcentrate, 12.3 g sucrose, 0.01 ml orange flavouring/100 ml water, sampleG = 12.5 g concentrate, 10 g sucrose, 0 ml orange flavouring/100 ml water, andsample H = 12.5 g concentrate, 10 g sucrose, 0.01 ml orange flavouring/100 mlwater.

K.A. Hein et al. / Food Quality and Preference 26 (2012) 35–44 41

3.3.3. Order of sample likingNo significant interaction was observed between the four black-

currant juice samples and the four contexts (F9,1012 = 0.995,p = 0.442). Results indicated that the blackcurrant juices were likedin a similar rank order among the three evoked contexts and con-trol context. Therefore, while blackcurrant juice samples were dis-criminated differently among the four contexts (see Section 3.3.1),there were no overall statistical differences in order of liking.Although not significantly different, mean ratings were lower inthe refreshing context compared to the control context. The great-est magnitude of difference in mean sample hedonic ratings wasseen between the control and movie contexts. The sample withboth added sugar and orange flavouring (F) was rated 1.0 pointlower, and the sample with added sugar (E) was rated 0.9 pointlower on the 9-point hedonic scale in the movie context comparedto the control context. From Fig. 3, it was also noted that the sam-ple with added orange flavouring (H) was rated as the least liked inthe refreshing context, breakfast context, and control context.However, the same sample (H) was rated as the most liked in themovie context.

3.4. Consumers’ perception

The four consumption contexts were found to be similar in howeasy/difficult it was for consumers to rate their liking/disliking ofthe samples (F3,253 = 2.03, p = 0.110), and consumers’ perceivedaccuracy of their liking information (F3,252 = 0.53, p = 0.665) (Table2). Therefore the use of an evoked context was not perceived byconsumers as impairing or improving their perceived performancefor the hedonic rating tasks.

4. Discussion

The aim of the present study was to examine how consumer he-donic responses for two different fruit beverages with subtle sen-sory differences are affected by different evoked consumptioncontexts in the laboratory test setting. Previous research has ex-plored consumer hedonic responses for products elicited in differ-ent eating locations (Edwards et al., 2003; King et al., 2004;Meiselman et al., 2000; Sverkén, Wendin, & Aström, 2009). In thisstudy the use of written scenarios to evoke different consumptioncontexts was explored, and the effects of these contexts on hedonicresponses in the laboratory test setting were investigated. This re-search offers consumer sensory researchers an approach that canbe used to provide consumers with a product’s consumption con-text or multiple different contexts when giving their hedonic re-sponses for products in the laboratory setting.

4.1. Were the three different consumption contexts evoked usingwritten scenarios?

To achieve the overall aim, it was first necessary to demonstratethat the three consumption contexts (while eating breakfast, whilewatching a movie, and while having something refreshing to drink)were evoked using the different written scenarios. The writtendescriptions provided by the consumers established that differentoccasions were imagined for each of the three evoked contexts.Although consumers imagined having a drink for a givenconsumption context, specific aspects of the context varied amongthe consumers. For example (Table 1), one consumer imagined‘being at café with their partner, eating a smoked salmon bagel

42 K.A. Hein et al. / Food Quality and Preference 26 (2012) 35–44

and drinking coffee and orange juice, while music played and thesun shined’. On the other hand, another consumer imagined ‘beingat home with their parents, eating a cooked breakfast with tea’.While consumers may have had different depictions of the evokedcontext, they imagined the intended context. The written descrip-tions indicated that consumers have different ideas as to what ‘de-fines’ their consumption context, demonstrating that context isindividualised. Unlike previous studies manipulating physical as-pects of a consumption context, written scenarios allowed con-sumers to personalise the evoked context.

While consumers were able to provide written descriptions ofthe evoked consumption contexts, differences in how the three con-texts were used by the consumer were found. Consumers in the mo-vie context, who were less compelled to desire something to drinkthan consumers in the breakfast and refreshing contexts, also indi-cated that the occasion imagined was not vivid in their minds whilethey evaluated the juices. Participants may have found that rating anapple or blackcurrant juice was difficult when imagining a movie oc-casion, as the majority of consumers described consuming a carbon-ated beverage. Consumers perceived both apple and blackcurrantjuice samples as being less appropriate to consume in the moviecontext than in the other two evoked contexts. Although pilot test-ing suggested that neither beverage would be inappropriate in theevoked contexts, this was determined in the absence of any actualproduct tasting. It may be that perceived beverage appropriatenessin the participant’s imagined occasion may change following tastingof products. Cardello and Schutz (1996) established that perceivedproduct appropriateness varies with use situation. Therefore it isimportant to consider the match between the product being evalu-ated, and the consumption context.

4.2. Effect of different evoked consumption contexts on hedonic rating

An objective of this study was to determine the effect of differ-ent evoked contexts (breakfast, movie and refreshing) on hedonicratings for two different fruit beverages (apple and blackcurrant).With the exception of the evoked movie context, the apple juicesamples were liked similarly in the evoked breakfast and refresh-ing contexts, and the control setting. Lower liking of blackcurrantjuice and greater discrimination of hedonic ratings in the evokedcontexts compared to the control context were observed.

While previous literature reported greater liking in a product’sconsumption context (Boutrolle, Delarue, Arranz, Rogeaux, &Köster, 2007; de Graaf et al., 2005; Meiselman et al., 2000), thepresent study demonstrated a decrease in liking for blackcurrantjuice in the evoked contexts. For example, the sample containingadded sugar and orange flavour (F) was rated 1.0 point lower onthe 9-point hedonic scale in the movie context, compared to thecontrol context. Whether a product is more or less liked in a con-sumption context can influence product-based decisions. Boutrolleand Delarue (2009) reported that French companies used 7, on a 1–10 scale, as an ‘action standard’ to decide on a product’s future inthe product development cycle. Meiselman, King, and Hottenstein(2004) have suggested that adding as much as 0.5–1.0 point of the9-point hedonic scale to a mean hedonic rating measured in a lab-oratory setting, better represented ratings obtained under real con-ditions (Cited in, Meiselman, 2008). However this studydemonstrated that depending on the consumption context, likingcould be lower. The present study further demonstrated that dif-ferent conclusions regarding the level of product liking may bedrawn when a product’s consumption context is not taken into ac-count. However this may also depend on the product type. This issimilar to King et al. (2004) who reported the effects of social inter-action, physical environment, food choice and the serving of prod-uct as a meal component on the liking of two variations of salad,pizza and iced tea were not consistent for the different food items

or between the variations of the food items. However consumptioncontexts may influence the liking of some products, but not others.Therefore the interaction between the product and context is crit-ical. One context may cause an effect for a product where no effectwas found for a different context.

Overall, evoked consumption contexts had a larger impact onthe hedonic ratings of blackcurrant juice than apple juice. Depend-ing on the type of product, differences in liking for samples elicitedin evoked contexts may also be associated with consumer aware-ness of the variation in sensory characteristics within the producttype. In general, a greater number of apple juice based products areavailable for purchase compared to blackcurrant juice based prod-ucts in New Zealand. Apple juice is commonly used as a back-ground to different fruit juices. Consumers may have been moreaccustomed to the wide variation in sensory characteristics ofcommercially available apple juice products and subsequently‘liked’ the juice samples regardless of the variation. Alternativelyconsumers may not have been familiar with variations in the sen-sory characteristics of blackcurrant juices and therefore had differ-ent hedonic responses to the blackcurrant juice samples,depending on the evoked context. While the effect of the evokedcontexts on hedonic response observed in blackcurrant juice com-pared to apple juice could be related to the order of juice presenta-tion, differences between contexts existed within juice type.

Evaluation of products in an evoked consumption context thatis inappropriate in relation to the product type, may have a nega-tive impact on hedonic ratings (Land, 1988). In the present study,when products were perceived as less appropriate in the evokedconsumption context, mean hedonic ratings were lower than whenelicited in the control context, and when elicited in an evoked con-text with greater product appropriateness. For example, lowermean hedonic ratings were observed in the evoked movie contextcompared to the other consumption contexts for both juice types,although not significantly in the case of apple juice. Blackcurrantjuice was not perceived as appropriate to drink at the movies,which may have been the basis for the samples to be less likedcompared to when no consumption context was evoked. It mayalso be the case that consumers are more critical of products whenthey do not ‘fit’ the consumption context. Therefore it is importantto consider the match between the product and context that isevoked. Further research is required to investigate the effect ofevoked contexts on hedonic responses when the context and prod-uct do not match.

While different effects of the evoked consumption contexts onmean hedonic responses, and ability to detect differences in sam-ple liking were observed for the apple and blackcurrant juices,there were no changes in sample preference within product typeamong the four consumption contexts. King et al. (2007) com-mented that context was unlikely to be an important considerationwhen designing a consumer test if changes in ranking of hedonicscores were of interest. However the comment was with respectto changes in hedonic responses to food components when manip-ulating the evaluation environment, social interaction, meal con-text and the consumer’s ability to choose the components for themeal. While no change in ranking of hedonic scores for the appleor blackcurrant juices was observed by evoking consumption con-texts in the present study, the authors consider it possible thatchanges in preference are possible depending on the consumptioncontext. For example, hot and cold chocolate beverages may beequally liked. However the hot chocolate beverage may be pre-ferred on a cold winter’s day while a cold chocolate beveragemay be preferred on a hot summer’s day. In this case, changes inconsumer product preference could depend on the consumptioncontext, and may need to be considered when designing a con-sumer test. However it is also important to consider the type ofsamples evaluated, and that the effect of context on hedonic re-

K.A. Hein et al. / Food Quality and Preference 26 (2012) 35–44 43

sponse may depend on the differences in type and size of sensorycharacteristics among the samples. Further research is needed tounderstand the changes in product liking and product preferencewithin different consumption contexts are evoked, and the influ-ence on different products varying in sensory characteristics.

Depending on the consumption context, consumers’ attentionmay be drawn to different product aspects, which may impact he-donic ratings. McEwan (1997) found that liking of product aspectssuch as appearance, flavour and texture varied depending onwhether products were evaluated in a central location or in-hometest (McEwan, 1997). Kozlowska and colleagues (2003) reportedthat less sweet apple juices were more liked in-home than in a lab-oratory setting. However the opposite trend was observed forsweeter samples. In the present study, the sweet blackcurrant juice(E) was liked to a greater extent than sample H in the evokedrefreshing context, compared to the breakfast context. Howeverin the evoked breakfast context, sample G, the non-sweet samplewas liked to a greater magnitude than H. Even more interestingis the observation that the sample with added orange flavour (H)was more liked in the movie context than the refreshing context,breakfast context and control context. This was seen by the direc-tion of the t-values that compared individual sample pairs (Fig. 4).Consumers may have responded to sensory cues more or less so, orto different ones, depending on the evoked context.

4.3. Consumers’ perception of task completion

The third objective of this study was to determine whether theevoked contexts influenced how easy or difficult it was for consum-ers to give hedonic responses for the products. Consumers in theevoked contexts found the task of hedonic ratings as easy as con-sumers in the control context. Consumers also perceived their likinginformation as equally accurate between the evoked contexts andthe control context. Hein et al. (2010) reported that consumers inan evoked refreshing context found hedonic ratings for a set ofapple juice samples to be as easy as in a control context when no con-sumption context was evoked. These results suggest that an evokedcontext does not increase the difficulty of eliciting hedonic ratingsfrom the consumer’s perspective. It is important to acknowledgethat consumer responses to questions regarding task completionin the evoked context may have been confounded with time and ef-fort required to describe their imagined occasions. Consumers inthe evoked contexts may have found less time and effort was re-quired for the hedonic task, following writing of the imagined con-sumption context. However, by having consumers imagine anddescribe a consumption context, consumers may have relaxed insome way and found the hedonic task easier than when a contextwas not evoked.

4.4. Additional considerations when evoking consumption contexts

In a study by Hein et al. (2010) differences in liking for the sameset of apple juices used in the present study were observed in anevoked refreshing context, but not in a controlled setting whenno consumption context was evoked. However it is important topoint out that semantic differences existed in the written scenariosused to evoke the refreshing context during the present study, andthat used by Hein et al. (2010). In the present study, consumerswere asked to imagine an occasion when they are ‘having’ arefreshing drink, while the study by Hein et al. (2010) asked con-sumers to imagine an occasion when they are ‘wanting’ a refresh-ing drink. As a result of semantic differences, consumers in thestudy by Hein et al. (2010) who evoked context of ‘wanting’ arefreshing drink, tended to described occasions of needing/wantinga beverage. Consumers in the present study who evoked context of‘having’ a refreshing drink described occasions when they were

actively consuming beverages, often times mentioning a specificbeverage. However, differences in written descriptions of evokedcontexts provided by consumers between the two studies areapparent. Therefore no difference in liking for the apple juicesmay be attributed to the change in wording of the written scenarioused to evoke the refreshing context.

‘Wanting’ a refreshing drink suggests a requirement to satisfy aphysiological need to quench thirst and has been described as ameasure of appetite (Havermans, Janssen, Giesen, Roefs, & Jansen,2009). For example, Beucler, Drake, and Foegeding (2005) reportedthat consumers consumed ‘thirst-quenching’ beverages when theywere thirsty (46%) and after exercise (34%). Consumers ‘wanting’ abeverage may have created an expectation (Cardello & Sawyer,1992) that the product would meet this physiological need to re-fresh them. On the other hand, ‘having’ a refreshing drink suggeststhat the requirement to satisfy one’s thirst is being met when tast-ing the test product, causing products to be liked equally regard-less of the differences in sensory characteristics. Further researchis needed to understand to what extent changes in wording influ-ence how consumption contexts are evoked and their effect on he-donic responses.

Written scenarios allow for diverse contexts to be evoked withinthe laboratory setting without requiring modification of the physi-cal environment. Written scenarios allow research flexibility in thetypes of consumption context that can be evoked and can providethe researcher with consumption contexts that may not otherwisebe possible to achieve. Therefore evoked context can be used to testthe effect of different contexts on products, and to test for interac-tion between contexts and products in a laboratory setting.

5. Conclusions

Written scenarios effectively evoked different consumptioncontexts in the laboratory setting. The elicitation of hedonic re-sponses of products was influenced by the evoked consumptioncontext, although the effects depended on the type of consumptioncontext and varied between product types. A greater effect ofevoked consumption contexts on hedonic responses was observedfor a set of blackcurrant juices, compared to a set of apple juices.When eliciting hedonic responses to products in evoked consump-tion contexts, it is important to consider the type of product, thetype of consumption context evoked, and how well matched thecontext and the product being investigated are. Written scenariosmay have an application to evoke consumption contexts in a labo-ratory setting to better understand product liking when testing inthe real context is not practical or possible to test in. Evoked con-text can allow the same product to be tested in different contexts,or different products to be tested in one context. Therefore, infor-mation regarding product and context interactions can be gainedby the use written scenarios to evoke context in the laboratory set-ting. By evoking consumption contexts in the laboratory setting,consumers may be provided with a frame of reference to providetheir hedonic responses without making the hedonic task more dif-ficult. While this research suggests that depending on the con-sumption context different sensory characteristics of a productmay be more or less important, further research is needed tounderstand how an evoked consumption context may impact theperception of specific sensory characters in relation to hedonic re-sponses. Research is also required to better understand the interac-tion between the consumption context being evoked and productcategory on consumer hedonic responses.

Acknowledgements

Authors would like to thank the staff at The New Zealand Insti-tute for Plant & Food Research Ltd., for help in piloting of the

44 K.A. Hein et al. / Food Quality and Preference 26 (2012) 35–44

written scenarios. Author K.A. Hein would like to acknowledgesupport from Rose Marie Pangborn Sensory Science Scholarship.

References

Bell, R., & Meiselman, H. L. (1994). Effects of adding an Italian theme to a restauranton the perceived ethnicity, acceptability and selection of foods. Appetite, 22,11–24.

Beucler, J., Drake, M. A., & Foegeding, E. A. (2005). Design of a beverage from wheypermeate. Journal of Food Science, 70, 277–285.

Bisogni, C., Falk, L. W., Madore, E., Blake, C., Jastran, M., Sobal, J., et al. (2007).Dimensions of everyday eating and drinking episodes. Appetite, 48, 218–231.

Boutrolle, I., & Delarue, J. (2009). Studying meals in the home and in the laboratory.In H. L. Meiselman (Ed.), Meals in science and practice. Interdisciplinary researchand business applications (pp. 128–165): Woodhead Publishing Ltd.

Boutrolle, I., Delarue, J., Arranz, D., Rogeaux, M., & Köster, E. (2007). Central locationtest vs. home use test: Contrasting results depending on product type. FoodQuality and Preference, 18, 490–499.

Cardello, A. V., & Sawyer, F. M. (1992). Effects of disconfirmed consumerexpectations on food acceptability. Journal of Sensory Studies, 7, 253–277.

Cardello, A. V., & Schutz, H. G. (1996). Food appropriateness measures as an adjunctto consumer preference/acceptability evaluation. Food Quality and Preference, 7,239–249.

de Graaf, C., Cardello, A. V., Kramer, F. M., Lesher, L. L., Meiselman, H. L., & Schutz, H.G. (2005). A comparison between liking ratings obtained under laboratory andfield conditions: The role of choice. Appetite, 44, 15–22.

Edwards, J. S. A., Meiselman, H. L., Edwards, A., & Lesher, L. L. (2003). The influenceof eating location on the acceptability of identically prepared foods. Food Qualityand Preference, 14, 647–652.

Havermans, R. C., Janssen, T., Giesen, J. C. A. H., Roefs, A., & Jansen, A. (2009). Foodliking, food wanting, and sensory-specific satiety. Appetite, 52, 222–225.

Hein, K. A., Hamid, N., Jaeger, S. R., & Delahunty, C. M. (2010). Application of awritten scenario to evoke a consumption context in a laboratory setting: Effectson hedonic ratings. Food Quality and Preference, 410–416.

Hersleth, M., Mevik, B. H., Naes, T., & Guinard, J. X. (2003). Effect of contextualfactors on liking for wine – Use of robust design methodology. Food Quality andPreference, 14, 615–622.

Hersleth, M., Ueland, O., Allain, H., & Naes, T. (2005). Consumer acceptance ofcheese, influence of different testing conditions. Food Quality and Preference, 16,103–110.

King, S. C., Meiselman, H. L., Hollenstein, A. W., Work, T. M., & Cronk, V. (2007). Theeffects of contextual variables on food acceptability: A confirmatory study. FoodQuality and Preference, 18, 58–65.

King, S. C., Weber, A. J., Meiselman, H. L., & Lv, N. (2004). The effect of meal situation,social interaction, physical environment and choice on food acceptability. FoodQuality and Preference, 15(7–8), 645–653.

Köster, E. P. (2003). The psychology of food choice: Some often encounteredfallacies. Food Quality and Preference, 14, 359–373.

Kozlowska, K., Jeruszka, M., Matuszewska, I., Roszkowski, W., Barylko-Pikielna, N., &Brzozowska, A. (2003). Hedonic tests in different locations as predictors ofapple juice consumption at home in elderly and young subjects. Food Qualityand Preference, 14, 653–661.

Land, D. G. (1988). Negative influences on acceptability. In D. M. H. Thomson (Ed.),Food acceptability (pp. 475–483). London: Elsevier.

Lawless, H. T., & Heymann, H. (1999). Sensory evaluation of food: Principles andpractices. Gaithersburg: Aspen Publishers, Inc.

MacFie, H. J. H., Bratchell, N., Greenhoff, K., & Vallis, L. (1989). Designs to balance theeffect of order of presentation and first-order carry-over effects in hall tests.Journal of Sensory Studies, 4, 129–148.

McEwan, J. A. (1997). A comparative study of three product acceptability trials. FoodQuality and Preference, 8, 183–190.

Meiselman, H. L. (1996). The contextual basis for food acceptance, food choice andfood intake: the food, the situation and the individual. In H. L. Meiselman & H. J.H. MacFie (Eds.), Food choice acceptance and consumption (pp. 239–263).London: Chapman & Hall.

Meiselman, H. L. (2008). Experiencing food products within a physical and socialcontext. In H. N. J. Schifferstein & P. Hekkert (Eds.), Product experience(pp. 559–580). San Diego, CA: Elsevier Ltd.

Meiselman, H. L., Johnson, J. L., Reeve, W., & Crouch, J. E. (2000). Demonstrations ofthe influence of the eating environment on food acceptance. Appetite, 35,311–327.

Meiselman, H. L., King, S., & Hottenstein, A. W. (2004). Laboratory product testingproduces an underestimation of true product acceptability. Florence, Italy: Paperpresented at the A Sense of Identity.

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:Does the drinking environment have any influence? Food Quality and Preference,18, 161–172.

Premavalli, K. S., Wadikar, D. D., & Nanjappa, C. (2009). Comparison of theacceptability ratings of appetizers under laboratory, base level and high altitudefield conditions. Appetite, 53, 127–130.

Sverkén, A., Wendin, K., & Aström, A. (2009). Comparison of home use test andcentral location test by the use of preference mapping. Paper presented at the8th Pangborn Sensory Science Symposium, 26–30th of July in Florence, Italy.