Appetite (2002) 39, 25±33doi:10.1006/appe.2002.0492, available online at http://www.idealibrary.com on

1

Research Report

Does information aboutMSG (monosodiumglutamate) content influence consumerratings of soupswith andwithoutaddedMSG?John Prescott and ArianeYoungSensory Science Research Centre,University of Otago, Dunedin,New Zealand

(Received19 October 2001, revision15March 2002, accepted inrevised form 22March 2002)

Responses to recent concerns about perceived adverse health effects of monosodium glutamate (MSG) haveincluded using prominent labels, e.g., `̀ No added MSG'', on products. Label information has been shown to createexpectations for a food's sensory properties and acceptability, and influence evaluations of the product. To assessthe impact of information about MSG content, subjects evaluated saltiness, richness, natural taste of, and liking for,vegetable soups with (MSG�) and without (MSGÿ) added MSG. Their attitudes to MSG were evaluated and foundto be generally negative. Subjects tasted both soups under three information conditions, presented as an ingredientlist: contains added MSG, does not contain added MSG, or no mention of MSG. The expected changes in liking andsensory properties due to added MSG were found, but there were no effects of information. A second experimentused a more obvious manipulation of information on added MSG content, as well as two soup flavors to reduce theobvious differences between the MSG� and MSGÿ soups. Again, there were no effects of information. These datasuggest that sensory properties are weighted more than information when products are evaluated during tasting,even when the information is highly relevant to beliefs and attitudes.

# 2002 Elsevier Science Ltd. All rights reserved.

Introduction

Glutamate, derived from glutamic acid, a non-essentialamino acid, is an important contributor to flavorin cuisines throughout the world. Sauces (soy;Worcestershire), cheese (especially Parmesan), toma-toes, mushrooms, or meat, fish and vegetable stocksadded to foods all have the effect of increasingglutamate levels (Ninomiya, 1998; Yamaguchi, 1998).Since the early 1900s, monosodium 1-glutamate(MSG) has been commercially manufactured for useas a flavor enhancer for the same purpose. There isample evidence that the addition of MSG (or relatedcompounds such as inosine monophosphate (IMP) and

Address correspondence to: J. Prescott, Sensory ScienceResearch Centre, University of Otago, PO Box 56, Dunedin,New Zealand. E-mail: john.prescott@stonebow.otago.ac.nz

0195±6663/02/$ ± see front matter

guanosine monophosphate (GMP)) to suitable foodshas a considerable impact on their sensory and hedonicproperties ± typically resulting in increased richness,savouriness, saltiness (because of the Na�) andpalatability (Bellisle et al., 1991; Fuke and Shimizu,1993; Prescott, 2001; Yamaguchi, 1998; Yamaguchi &Takahashi, 1984).

During the past three decades, there has been sub-stantial controversy regarding the use of MSG in foods,at least in Western countries. The original source of thiscontroversy appears to be a letter to the New EnglandJournal of Medicine (Kwok, 1968) in which it wasspeculated that MSG (as one alternative among severalother ingredients) could be the cause of adverse reac-tions following consumption of Chinese restaurantfood reported by the author and others. This article andsubsequent publicity about MSG seems to have tappedinto more general consumer concerns regarding food

# 2002 Elsevier Science Ltd. All rights reserved.

26 J. Prescott and A.Young

additives, resulting in an increasingly widespread beliefamong consumers that MSG is responsible for allergicreactions ± variously, asthma or a `̀ Chinese restaurantsyndrome'' of numbness, weakness, headaches andpalpitations. However, several placebo-controlled stu-dies have failed to find evidence implicating MSG as acause of these or other adverse reactions (Geha et al.,2001; Tarasoff & Kelly, 1993). Considerations of boththe clinical and laboratory evidence forMSGproducingadverse reactions have concluded that it is a safe foodingredient, a status accepted by both the US Food &Drug Administration and Institute of Food Technolo-gists (IFT, 1987; Walker & Lupien, 2001).

Nevertheless, and despite the fact that MSG is oftenrequired to be included among the list of food ingre-dients, many food manufacturers have increasinglyadopted a strategy of placing additional prominentmessages regarding MSG on food labels. As a result,food labels advertising `̀ No added MSG'' have becomecommonplace (Dillon, 1993). For example, a recentsurvey (unpublished) in Australian supermarkets foundthat, on average, negative claims such as `̀ No addedMSG'' were present on 59% of canned soups and 33%of convenience meals such as noodle and pasta dishes.One possible consequence of such labels is that they

generate and reinforce beliefs thatMSG is harmful and/or an unsafe ingredient. Recent research on the effects ofdifferent types of label information suggests the possi-bility that these messages may also influence theacceptability of products containing added MSG. BothSchutz and Lorenz (1976) and Johansson et al. (1999)examined the impact of information about growingmethod on consumer preferences for organic and con-ventional vegetables. In both studies, relative to thesesame foods unlabelled, products labelled as organicgenerally showed increases in measures of preference.Food acceptability has also been shown to be influencedby information about nutritional qualities (Kahkonenet al., 1996), while information on food ingredientsand use (Tuorila et al., 1994a), taste (Pelchat & Pliner,1995) or nutritional (McFarlane & Pliner, 1997) quali-ties have produced increases in the willingness to trynovel foods.

These effects extend to the perception of sensoryqualities. Kahkonen and Tuorila (1998) found thatwhenBologna sausageswere labelled as reduced fat, theywere rated as lower in saltiness and fattiness than aregular sausage, differences that were not apparentwhen no information was given. Similar effects on rat-ings of fattiness/leanness, quality, and greasiness ofsamples of beef were seen by Levin and Gaeth (1988)even though they manipulated only the emphasis(framing) ± 25% fat or 75% lean ± and not the contentof the information.

Themost likely explanation for the impact of labels onhedonic and sensory responses to foods is through thecreation of expectations (Deliza & MacFie, 1996). In ademonstration of the impact of expectations created bylabelling on product evaluation, Tuorila et al. (1994b)found clear differences in the expected liking and sen-sory qualities between fat-free and regular fat versionsof several foods ± differences that were much greaterthan those apparent when the foods were actuallytasted. They showed also that when these products werelabelled, relative to an unlabelled condition, hedonicand sensory responses to the actual product weremade more congruent with those that were expected.Similar examples of `̀ assimilation'' of ratings of actualproducts with expected ratings have been reported bySchifferstein et al. (1999) and Cardello and Sawyer(1992).

Label informationmay be especially effective in thosewhose prior attitudes are congruent with that informa-tion. A common finding is that sensory/acceptabilityjudgements appear to be brought into line with priorbeliefs/attitudes. Thus, high fat users rated a high-fatproduct as closer to ideal when it was labelled as such,relative to blind tasting (Daillant-Spinnler & Issanchou,1995). Aaron et al. (1994) demonstrated that subjectswho had positive attitudes towards reduced fat spreadsrated such a product asmore pleasant, more spreadable,and with better mouthfeel than did those whose atti-tudes were more positive towards full fat spreads.Dividing subjects according to beliefs showed similaradditional effects for smoothness & flavor.

The present studywas designed to examine the impactof information that specified whether MSG was addedto foods or not on ratings of the hedonic and sensoryproperties of soups. In particular, we wished to know ifsuch information would modify ratings of the sensoryand hedonic impact that the actual addition ofMSGhason the soup. Thus, we manipulated orthogonally theactual presence of MSG in the soups and the informa-tion provided about its presence. In addition, we mea-sured beliefs and attitudes towardsMSG in foods with aview to providing a context within which any effectscould be interpreted.

Experiment1

MethodsMethods

Pilot study

Since we wished to examine the impact of labellingrelative to the changes in acceptability and sensoryproperties produced by added MSG, a pilot study wasconducted to assess the impact of different levels of

MSGand consumer ratings of soups 27

added MSG. This would also allow us to select asuitable level of MSG to add to the soups used in themain study.

Twenty-five subjects (3 males and 22 females; meanage� 22 years; range: 20±24 years), all universitystudents, tasted five samples of Continental brandVegetable Stockpot soup, which claimed to have noadded MSG in its ingredients. The five samples con-sisted of a base sample of soup, made up from a dry mixwith water according to manufacturer's instructions, towhich was added MSG at 0�4%, 0�8%, 1�2% and 1�6%w/w. Subjects tasted and swallowed 25ml of each of thesamples, which were presented one at a time, and in abalanced order. The samples were rated for overall liking(dislike extremely to like extremely), saltiness, richnessand natural taste (low to high) using 100-mm line scales.

Repeated measures ANOVA (SPSS V. 10), followedby post-hoc testing using Fisher's LSD test, revealedthat the addition of MSG increased overall liking(F(4,96)� 5�34, p< 0�005), richness (F(4,96)� 6�61,p< 0�001), and saltiness (F(4,96)� 5�5, p< 0�001). Ineach case, the addition of any level ofMSG significantlyincreased ratings. The addition of MSG failed to influ-ence ratings of natural taste (F(4,96)� 0�91, p� 0�46).A level of 0�8% MSG was selected for the main study,since this concentration is within recommended usagelevels for a variety of different foods.

Main study

Subjects. Sixty subjects (20 males and 40 females; meanage� 34 years; range: 20±68 years) were recruitedfrom university staff, students and the general public.

SamplesandMSGinformation. Samples of the same soupused in the pilot study were prepared according tomanufacturer's instructions, split into two batches, andthen frozen until required. Prior to freezing, 0�8% w/wMSG was added to one batch. Prior to serving,samples were thawed overnight, and then heated to70 �C.

The samples were presented to subjects under threeinformation conditions: (1) no information; (2) thestatement: `̀ No addedMSG (monosodium glutamate)'';and (3) the statement: `̀ Added MSG (monosodiumglutamate)''. The design was fully factorial, with eachsample type (0�8% MSG added: MSG�; no addedMSG: MSGÿ) combined with each label condition tocreate six samples in total.

The label information was presented to subjects via asoup ingredient list that was presented immediatelyprior to tasting each sample. The subjects were told thata soup manufacturer, in addition to wanting informa-tion on the sensory properties and liking for different

versions of the soup,wished to knowwhether consumerswere familiar with the soup's ingredients. The list wasbased on the soup's actual ingredients ± ten items, listedfrom highest to lowest percentage of content ± togetherwith two additions. The first of these was the informa-tion regarding MSG, which was presented as the lastingredient on the list. In addition, so that this latterinformation did not alert subjects to the real purpose ofthe research, two other ingredients (from a list ofcommon herbs such as rosemary, thyme, oregano,cumin, basil) were also added to the list and varied fromsample to sample. To ensure that subjects read each itemon the list, they were asked to indicate their degree offamiliarity by circling one of three options ± familiar,unfamiliar, and unsure ± that were next to each ingre-dient. Only those subjects who indicated that they werefamiliar with MSG were included in the data analysis.

Procedure. Subjects attended two sessions, tasting threeof the six samples in each session. Sample presentationorder was determined using a Williams Latin Squarebalanced for first order and carry-over effects (MacFieet al., 1989). Prior to receiving each sample, subjectscompleted the relevant ingredient checklist. Thesamples (25ml), served one at a time, were tastedand swallowed, and then rated for overall liking (dislikeextremely to like extremely), saltiness, richness andnatural taste (low to high) using 100mm line scales.

In order to characterise subjects' attitudes towardsMSG in foods, a 24-item questionnaire was adminis-tered at the completion of all other data collection. Thequestionnaire contained a variety of items related to saltand fat in foods, geneticallymodified and organic foods,reading ingredient labels, aswell as five items concerningMSG in foods. These latter items were based oncommon beliefs regarding MSG, and its health andsensory impact. The majority of items were included asdistracters, and only the MSG-related responses wereanalysed. These items consisted of the followingstatements:

± MSG (monosodium glutamate) improves the flavorof foods

± It is better to choose foods without added MSG(monosodium glutamate)

± Eating foods with MSG (monosodium glutamate) isunhealthy

± Foods with added MSG (monosodium glutamate)taste less natural

± MSG (monosodium glutamate) in foods can causeallergic reactions in some people

In each case, subjects rated their agreement withthese statements using a scale with the following seven

28 J. Prescott and A.Young

categories (scored 1±7): disagree strongly, disagreemoderately, disagree slightly, neither agree nor disagree,agree slightly, agree moderately, and agree strongly.

Analysis. Ratings were converted into scores from 0 to100 and analysed using repeated measures ANOVA(SPSS V. 10), with Information (no information; `̀ Noadded MSG''; `̀ Added MSG'') and Added MSGcontent (MSG+ /MSGÿ) as within-subject factors.

ResultsResults

Data from two subjects were removed from the studydue to their failure to complete both test sessions.Responses from seven other subjects were not includedfor analysis because they indicated in the ingredientchecklist that they were unfamiliar with MSG.

MSGattitudes/beliefs

On the whole, based on the responses to each of theMSG-related questionnaire items, the subjects heldnegative attitudes/beliefs towards MSG in foods.Although there was a tendency to agree that MSGimproved the flavorof foods (median� 5/agree slightly),

Figure1. Mean (�SE) ratings by subjects in Exp. 1 of overallMSGÿ samples under the three information conditions.

there was also agreement with the statements that:

± It is better to choose foods without added MSG(70% of responses were ratings of 5 or higher;median� 6/agree moderately);

± Foods with added MSG are unhealthy (66%;median� 6); and

± MSG can cause allergic reactions in some people(63%; median� 6).

The median response to Foods with added MSG tasteless natural was 4 (neither agree nor disagree),suggesting uncertainty regarding such effects.

Effects of MSG content & information

The mean ratings for each attribute for the differentMSG content and information conditions are shownin Fig. 1.

ANOVArevealed that the addition ofMSG increasedratings of overall liking (F(1,50)� 28�47, p< 0�001),richness (F(1,50)� 48�55, p< 0�001), saltiness (F(1,50)�59�07, p< 0�001), and natural taste (F(1,50)� 11�37,p< 0�005).

By contrast, there were no significant effects of thedifferent information conditions on overall liking(F(2,100)� 0�313, p� 0�73), richness (F(2,100)� 0�886,p� 0�42), saltiness (F(2,100)� 0�36, p� 0�7), and natural

liking, richness, saltiness and natural taste of the MSG+ and

MSGand consumer ratings of soups 29

taste (F(2,100)� 1�234, p� 0�3), and no significantinformation�MSG content interactions.

DiscussionDiscussion

The impact of the addition of 0�8% MSG to the soupwas consistent with previous research showing changesin both liking and sensory properties (Prescott, 2001;Yamaguchi, 1998). The generally negative attitudes ofthese subjects towards the use of MSG in foods led tothe expectation that information indicating that asoup contained added MSG would influence overallliking, at least. Despite this, the information manip-ulation failed to influence any ratings.

While it is clearly conceivable that consumers areuninfluenced by this type of labelling, particularly in thecontext of tasting foods, it is also possible that, in thepresent study, the informationmanipulation failed. Twopotential explanations for this suggest themselves. First,it may be that the method of presenting information byhaving subjects read an ingredient list containing theMSG information was ineffective. This may be becauseit was either too `̀ subtle'' a method of delivery, with theMSG information lost in the `̀ noise'' of other ingre-dients, or simply regarded by the subjects as a separatetask, unconnected to the subsequent tasting of thesample. Second, there were considerable differences inthe sensory properties of the soups with and withoutadded MSG. This may have undermined the believ-ability of the information, especially in the conditions inwhich there was a mismatch between label claim andactual MSG content.

To address both of these issues, we conducted asecond, similar, experiment but with two important dif-ferences from Experiment 1. The first of these was thatwe used two different soup flavors, both of which werepresented in plain and added MSG forms. For eachsubject, we deliberately confounded label informationwith flavor. Across all subjects, however, we balancedwhich soup flavor was associated with which type ofinformation. This meant that another sensory propertyapart from MSG content was associated with MSGinformation, while still allowing evaluation of therelative influence of information and MSG content.

The other major change was in the presentation of theinformation regarding MSG content, which was placedon the top of the response sheet thatwas used formakingratings. It is appreciated that this approach may be lesstrue to life than the previous method of delivering infor-mation via a product ingredient list. Nevertheless, it wasthought to be important, in assessing the potentialimpact of information, to make the message as obviousas possible, and to connect it directly to the sample to betasted.

Experiment 2

MethodsMethods

Pilot study

A pilot study was undertaken to ensure the two soups± Vegetable Stockpot and Minestrone soups, bothContinental brand ± were approximately equivalentlyliked and similarly influenced by the addition of MSG.Seventeen subjects (4 males and 13 females; meanage� 24 years; range: 19±32 years), all universitystudents, participated. These subjects evaluated foursamples ± two samples of each soup flavor, with andwithout added 0�8% w/w MSG. Subjects tasted andswallowed 25ml of each of the samples, which werepresented one at a time in a balanced order. Thesamples were rated for overall liking (dislike extremelyto like extremely), saltiness, richness and natural taste(low to high) using 100-mm line scales.

Repeated measures ANOVA (SPSS V. 10) withMSGcontent and Soup type as within-subject factors revealedthat the addition of MSG increased overall liking(F(1,16)� 25�29, p< 0�001), richness (F(1,16)� 12�23,p< 0�005), and saltiness (F(1,16)� 8�25, p< 0�05), butthere was no effect for ratings of natural taste(F(1,16)� 2�39, p� 0�14). There were no main effects orinteractions involving the different soup types, indicat-ing that each soup flavor was similar on the measuredattributes and had been similarly influenced by theaddition of MSG.

Main study

Subjects. Forty-six subjects (25 males and 21 females;mean age� 22 years; range: 21±26 years) wererecruited from university students and the generalpublic.

Samples and information. Samples of the same soupsused in the pilot study were prepared according tomanufacturer's instructions, and then frozen in twobatches per soup flavor until required. Prior tofreezing, 0�8% w/w MSG was added to one batch ofeach soup. Prior to serving, soup samples were thawedovernight, and then heated to 70 �C.

These samples were presented to subjects under twoinformation conditions: (1) `̀ No added MSG (mono-sodium glutamate)''; and (2) `̀ Added MSG (mono-sodium glutamate)''. This information was included inthe headings on the response form as part of thedescription of the soup tobe tasted, e.g.,Minestronewithadded MSG (monosodium glutamate).

30 J. Prescott and A.Young

Study design. For both soup flavors, each sample type(MSG+ /MSGÿ) was combined with each informa-tion condition to create four samples in total.However, in order to ensure that all subjects receivedboth soups, for each subject, one soup flavor wasassociated only with correct information/content pairs(`̀ No added MSG''/MSGÿ; `̀ Added MSG''/MSG+ )while the other flavor was associated only withincorrect combinations (e.g. `̀ No added MSG''/MSG�; `̀ Added MSG''/MSGÿ). This ensured thateach subject received the different flavors with bothMSG conditions. Hence, sensory differences betweensamples could not be unequivocally associated withMSG label information by the subject. To avoid soupflavor producing an overall confound with informationor MSG content, half of the subjects were allocated toreceive the Vegetable Stockpot soup with correctcombinations and the Minestrone with incorrectcombinations, and for the other half it was vice versa.

Procedure. Subjects attended a single session, in whichthey evaluated all four samples. Sample presentationorder was determined using a Williams Latin Squarebalanced for first order and carry-over effects (MacFieet al., 1989). The 25-ml samples, served one at a time,were tasted and swallowed, and then rated for overallliking (dislike extremely to like extremely), saltiness,richness and natural taste (low to high) using 100-mmline scales. In addition, subjects also rated purchaseintent using a line scale on which they indicated theirlevel of agreement (disagree strongly ± agree strongly)with the statement: Based on this sample, I wouldcertainly purchase this product.

The same questionnaire used in Exp. 1 was adminis-tered at the completion of all other data collection inorder to characterise these subjects' attitudes/beliefsregarding MSG in foods.

Analysis. Ratings were converted into scores from 0 to100 and analysed using repeated measures ANOVA(SPSS V. 10) with Information (No added MSG;Added MSG) and Added MSG content (MSG�/MSGÿ) as within-subjects factors. Because it wasbalanced across subjects, the association of thedifferent soup flavors with the different MSG con-tent/label combinations was not included as a factor inthe analysis.

ResultsResults

Attitudes towardsMSG

Again, there was a tendency for subjects to holdnegative attitudes/beliefs towards MSG, although it

was less pronounced than in Exp. 1, mainly due tohigher levels of uncertainty for several of the questions.For example, 39�1% and 69�6% of subjects, respec-tively, indicated that they neither agreed nor disagreedwith the statements regarding natural taste and allergicreactions. The percentage of subjects scoring 5 orhigher and the median scores for each item are asfollows:

± MSG (monosodium glutamate) improves the flavorof foods (56.5%; median� 5/agree slightly)

± It is better to choose foods without added MSG(monosodium glutamate) (60�9%; median� 5)

± Eating foods with MSG (monosodium glutamate) isunhealthy (47�8%; median� 4/Neither agree nordisagree)

± Foods with added MSG (monosodium glutamate)taste less natural (39�1%; median� 4)

± MSG (monosodium glutamate) in foods can causeallergic reactions in some people (30�4%;median� 4)

Effects of MSG content and information

The mean ratings for each attribute for the differentMSG content and information conditions are shown inFig. 2.

ANOVA revealed that soups containing added MSGhad higher ratings of overall liking (F(1,45)� 46�82,p< 0�001), richness (F(1,45)� 25�26, p< 0�001), saltiness(F(1,45)� 22�83, p< 0�001), natural taste (F(1,45)�5�91, p< 0�05), and purchase intent (F(1,45)� 28�33,p< 0�001).

As in Exp. 1, there were no significant effects of thedifferent information conditions for overall liking(F(1,45)� 0�001, p� 0�98), richness (F(1,45)� 0�02,p� 0�89), saltiness (F(1,45)� 2�08, p� 0�16), and naturaltaste (F(1,45)� 2�03, p� 0�16), and no impact onpurchase intent (F(1,45)� 1�59, p� 0�21). There were nosignificant information�MSG content interactions.

Discussion

Both studies reported here showed characteristiceffects of MSG on the sensory and hedonic propertiesof the soups ± increases in liking, richness, saltiness,and natural taste. In neither experiment, though, werethere any effects of the manipulation of informationabout MSG content. It is unlikely that the changes inthe design of the second experiment, made to ensurethat the subjects were aware of the labels, wereunsuccessful. The subjects were provided with veryobvious soup labels, which included a description ofMSG content. Moreover, it could be expected thatgiven their predominantly negative attitudes/beliefs

Figure 2. Mean (�SE) ratings by subjects in Exp. 2 of overall liking, richness, saltiness, natural taste of, and purchase intentfor the MSG� and MSGÿ samples under the two information conditions.

MSGandconsumer ratings of soups 31

towards the use MSG in foods, these subjects wouldattend to such information.

The question of how the subjects were classified ispertinent to this issue if, for example, the responses to thequestionnaire items overestimated the degree of nega-tive attitudes/beliefs held. The questionnaire items weredeveloped to reflect a range of beliefs commonly seen inrelation to MSG in foods. As a result, four of the fiveitems were negative statements ± MSG is unhealthy, lessnatural, produces allergic reactions, and foods withoutadded MSG should be chosen. The possibility exists,therefore, that the way the itemswere framed could haveinduced some bias in responding, although there is noway to determine this at this stage. The fact that theseitems were randomly embedded in a much larger ques-tionnaire covering several food-related topics may havecountered this potential bias, however, by taking focusoff MSG as a issue.

A clearer picture of the role of prior attitudes towardsMSG in influencing responses to food labels might beobtained in future studies by using a larger sample ofconsumers, and covering a greater range of such atti-tudes, from positive to highly negative. Given theskewed distributions of responses in the present experi-ments, the sample sizes used are unlikely to providesufficient statistical power for such an analysis.Apart from these considerations, the most parsimo-

nious explanation for the failure to find information

effects in either experiment is that the messages them-selves were ineffective in influencing ratings. What isparticularly surprising is that, in addition to the sensoryand hedonic ratings, ratings of purchase intent wereunaffected by information about MSG. It might beexpected that subjects could separate the issue of likingfor a sample and whether they would purchase a likedproduct, since the latter should be more strongly influ-enced by their attitudes towards other aspects of theproduct, such as perceived unhealthy ingredients. Whatthis implies is that responses to the actual sensoryproperties were weighted far more strongly that otheraspects of the product experience.Previous studies have also suggested that the impact

of information on sensory or hedonic ratings of foodsmay be a function of if, and when, the foods are tasted.Levin and Gaeth (1988) found that their labelling/framing effect was reduced when subjects actuallysampled meat products, relative to no sampling. More-over, information effects were greatest when the meatwas not tasted, next largest when label information wasgiven prior to tasting, and smallest when the meat wastasted prior to receiving information. Similarly, Kah-konen et al. (1996) noted that information on fat and saltcontent presented after home use of a low fat, low saltspread did not effect sensory or hedonic ratings, whereasthe same nutritional information presented beforehome use increased both pleasantness and melting-rate

32 J. Prescott and A.Young

ratings. In the present studies, the information wasostensibly presented prior to the tasting. However, theevaluations of the products, at which point the subjectsmight be expected to integrate information from bothextrinsic and intrinsic sources, took place after tasting.

One of the possible implications of such findings isthat while labels may be highly effective in creatingexpectations for products that are novel, andmay in factencourage an initial trial of such products (e.g. Tuorilaet al., 1994a), theymaybe relatively ineffective for repeatpurchase consumers when sensory properties may beparamount.

Information/label effects might also be expected tobe most effective when responses to products, orproduct differences, are equivocal or moderate inintensity. Hoch and Ha (1986) have suggested thatinformation framing is effective primarily when theproduct experience is ambiguous ± this might be thecase with novel foods, for example. In the present case,there was little ambiguity, since the actual differencesbetween the MSG+ and MSGÿ soups were soapparent, despite an attempt to vary other aspects ofthe soups in Exp. 2. Because of this, the currentexperiments may have made the sensory impact ofMSG much more salient than it usually is, perhapsundermining the label information.

Despite the lack of effects here, negative claimsregarding MSG on labels may still be influential duringproduct purchase. Consumers may appreciate, in ageneral sense, that MSG can enhance flavor (as indi-cated by the questionnaire responses). Yet, in contrastto the procedures in these studies, they are unlikely tomake direct taste comparisons of similar products withand without MSG claims. Moreover, since it is pri-marily a marketing device, the meaning of the phrase`̀ No added MSG'' is both highly ambiguous andunlikely to accurately reflect either the actual gluta-mate content of the food or the food's sensory prop-erties. Much of the time, therefore, consumers may beunable to make choices regarding MSG content basedon sensory properties.

With the existing negative attitudes that manyconsumers possess, and given the choice betweensimilar products, one of which has a negative MSGclaim, they may treat this claim as a positive feature.What is less clear is whether the absence of such claimson products is problematic. Clearly, given the highincidence of negative MSG claims, food manufacturersare adopting the view that they might be. Empiricalinvestigations of these issues would be worthwhile,perhaps using descriptions of novel products with andwithout negative MSG claims for which consumersrated their expected liking and expected sensoryproperties.

References

Aaron, J. L., Mela, D. J. & Evans, R. E. (1994). The influ-ences of attitudes, beliefs and label information on per-ceptions of reduced-fat spread. Appetite 22, 25±37.

Bellisle, F., Monneuse, M. O., Chabert, M., Laure-Achagiotis, C., Lanteaume, M. T. & Louis-Sylvestre, J.(1991). Monosodium glutamate as a palatability enhan-cer in the European diet. Physiology & Behavior 49,869±874.

Cardello, A. V. & Sawyer, F. M. (1992). Effects of discon-firmed consumer expectations on food acceptability.Journal of Sensory Studies 7, 253±277.

Daillant-Spinnler, B. & Issanchou, S. (1995). Influenceof label and location of testing on acceptability ofcream cheese varying in fat content. Appetite 24,101±106.

Deliza, R. & MacFie, H. J. H. (1996). The generation of sen-sory expectation by external cues and its effect on sensoryperception and hedonic ratings: A review. Journal ofSensory Studies 11, 103±128.

Dillon, P. M. (1993). Invasion of the MSG-free ingredients.Food Engineering 64, 133±134, 136.

Fuke, S. & Shimizu, T. (1993). Sensory and preferenceaspects of umami. Trends in Food Science & Technology4, 246±251.

Geha, R. S., Beiser, A., Ren, C., Patterson, R., Greenberger,P. A., Grammer, L. C., Ditto, A. M., Harris, K. E.,Shaughnessy, M. A., Yarnold, P. R., Corren, J. &Saxon, A. (2001). Review of alleged reaction to monoso-dium glutamate and outcome of a multicenter double-blind placebo-controlled study. Journal of Nutrition 130,1058S±1062S.

Hoch, S. J. & Ha, Y. W. (1986). Consumer learning: Adver-tising and the ambiguity of product experience. Journal ofConsumer Research 13, 221±233.

Institute of Food Technologists Expert Panel on Food Safetyand Nutrition (1987). Monosodium Glutamate (MSG):A Scientific Status Summary.

Johansson, L., Haglund, A., Berglund, L., Lea, P. &Risvik, E. (1999). Preference for tomatoes, affected bysensory attributes and information about growth condi-tions. Food Quality & Preference 10, 289±298.

Kahkonen, P., Tuorila, H. & Rita, H. (1996). How informa-tion enhances acceptability of a low-fat spread. FoodQuality & Preference 7, 87±94.

Kahkonen, P. & Tuorila, H. (1998). Effect of reduced-fat in-formation on expected and actual hedonic and sensoryratings of sausage. Appetite 30, 13±23.

Kwok, R. H. M. (1968). Chinese-restaurant syndrome. NewEngland Journal of Medicine 278, 796.

Levin, I. P. & Gaeth, G. J. (1988). How consumers are af-fected by the framing of attribute information beforeand after consuming the product. Journal of ConsumerResearch 15, 374±378.

McFarlane, T. & Pliner, P. (1997). Increasing willingness totaste novel foods: Effects of nutrition and taste informa-tion. Appetite 28, 227±238.

Macfie, H., Bratchell, N., Greenhoff, K. & Vallis, L. V.(1989). Designs to balance the effect of order of presenta-tion and first-order carry-over effects in hall tests. Journalof Sensory Studies 4, 129±148.

Ninomiya, K. (1998). Natural occurrence. Food ReviewsInternational 14, 177±211.

MSGandconsumer ratings of soups 33

Pelchat, M. L. & Pliner, P. (1995). `̀ Try it. You'll like it.''Effects of information on willingness to try novel foods.Appetite 24, 153±166.

Prescott, J. (2001). Taste hedonics and the role of umami.Food Australia 53, 550±554.

Schifferstein, H. N. J., Kole, A. P. W. & Mojet, J. (1999).Asymmetry in the disconfirmation of expectations fornatural yogurt. Appetite 32, 307±329.

Schutz, H. G. & Lorenz, O. A. (1976). Consumer preferencesfor vegetables grown under `̀ commercial'' and `̀ organic''conditions. Journal of Food Science 41, 70±73.

Tarasoff, L. & Kelly, M. F. (1993). Monosodium L-Gluta-mate: A double-blind study and review. Food & ChemicalToxicology 31, 1919±1035.

Tuorila, H., Meiselman, H. L., Bell, R., Cardello, A. V. &Johnson, W. (1994a). Role of sensory and cognitive

information in the enhancement of certainty andliking for novel and familiar foods. Appetite 23,231±246.

Tuorila, H., Cardello, A. V. & Lesher, L. L. (1994b). Antece-dents and consequences of expectations related to fat-freeand regular-fat foods. Appetite 23, 247±263.

Walker, R. & Lupien, J. R. (2001). The safety evaluation ofmonosodium glutamate. Journal of Nutrition 130, 1049S±1052S.

Yamaguchi, S. (1987). Fundamental properties of umamiin human taste sensation. In Kawamura, Y. andKare, M. R. (eds) Umami: A Basic Taste, N. Y., MarcelDekker, pp. 41±73.

Yamaguchi, S. & Takahashi, C. (1984). Interactions ofmonosodium glutamate and sodium chloride on salti-ness and palatability. Journal of Food Science 49, 82±85.