Food Quality and Preference 21 (2010) 394–401

Contents lists available at ScienceDirect

Food Quality and Preference

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

The application of check-all-that-apply (CATA) consumer profiling topreference mapping of vanilla ice cream and its comparison to classicalexternal preference mapping

Lauren Dooley, Young-seung Lee, Jean-François Meullenet *

Department of Food Science, University of Arkansas, 2650 N. Young Avenue, Fayetteville, AR 72704, United States

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

Article history:Received 19 January 2009Received in revised form 27 August 2009Accepted 11 October 2009Available online 21 October 2009

Keywords:Check-all-that-applyPreference mappingVanilla ice creamMultiple factor analysis

0950-3293/$ - see front matter � 2009 Elsevier Ltd. Adoi:10.1016/j.foodqual.2009.10.002

* Corresponding author. Tel.: +1 7853418710.E-mail address: jfmeull@uark.edu (J.-F. Meullenet)

This study was conducted to evaluate the use and efficacy of check-all-that-apply (CATA) data for the cre-ation of preference maps, and to compare these maps to classical external maps generated from tradi-tional sensory profiles. Ten commercial vanilla ice cream products were presented to 80 consumers.Consumers answered an overall liking question using the 9-point hedonic scale and a CATA question with13 attributes which described the sensory characteristics of vanilla ice cream. A trained descriptive panelof 17 individuals developed a profile of 23 attributes for the vanilla ice cream products. Preference mapscreated by CATA counts were compared to those by descriptive profiles via multiple factor analysis(MFA). The characterization of the products by both sensory methods showed very good agreementbetween the methods. The MFA of map configurations showed fair agreement between the techniquesused to produce the preference maps, indicating that CATA data applied to preference mapping gave sim-ilar results to external preference mapping.

� 2009 Elsevier Ltd. All rights reserved.

1. Introduction

Check-all-that-apply (CATA) questions regarding consumer-perceived product attributes have been used in consumer studiesto determine what sensory attributes may be characteristic of aspecific product (Lancaster & Foley, 2007). Some researchers al-ready advocate the use of consumer sensory profiling to lead prod-uct development as an alternative to classical sensory profiling(Punter, 2008; Worch, Lê, & Punter, 2008). The format of the CATAquestion allows consumers to choose all potential attributes fromthe given lists to describe the test products. This is different fromscaling in the sense that no intensities are given to the attributes.In addition, the descriptors are not constrained to product sensoryattributes, but could also be related to product usage or concept fit.This type of methodology has the advantage of gathering informa-tion on perceived product attributes without requiring scaling,allowing for a slightly less contrived description of the main sen-sory properties of the product tested (depending on how the termsare created).

The actual generation of CATA terms can be performed in manyways: the consumers can choose words to describe the productduring the test (modified free choice profiling), terms can be given

ll rights reserved.

.

by a trained panel, or terms can be generated by consumers nottesting the product (i.e. a focus group). Free choice profiling allowsconsumers to use as many or as few words as necessary to describethe product and evaluate the intensities of the chosen attributes,resulting in a less expensive and more accurate view of consumerperception and acceptance (Deliza, Macfie, & Hedderley, 2005;Gonzáles-Tomás & Costell, 2006; González Viñas, Garrido, & Wittigde Penna, 2001; Williams & Langron, 1984). However, if each con-sumer selects his/her own terms, the analysis becomes cumber-some since each term must be subjectively interpreted andcombined with similar terms (Meilgaard, Civille, & Carr, 2007).Seo, Lee, and Hwang (2009) used consumers to describe sensorycharacteristics of coffee. Verification of the terms was then con-ducted by other consumers to confirm that the terms were appro-priate and understandable. While this is an effective method, thetime required is to complete the test is extensive.

Terms generated by a trained panel have the benefit of beingmore comprehensive and better described, though they may betoo complex for the average consumer to understand and could re-quire simplification. Altering the terms in this manner is difficult todo while retaining the correct term description and definition.However, it has been shown that differences in sensory evaluationsbetween trained and untrained (naïve consumers) are minimal(Benedito, Cárcel, & Mulet, 2001; Guerrero, Gou, & Arnau, 1997;Husson & Pagés, 2003; Lelievre, Chollet, Abdi, & Valentin, 2008),so using less obscure terms by a descriptive panel could be a

L. Dooley et al. / Food Quality and Preference 21 (2010) 394–401 395

beneficial tool for creating a CATA list. Ultimately, it is the re-searcher’s decision as to which method is most appropriate.

The CATA method requires minimal instruction, is relativelyeasy to perform and is completed quickly (Lancaster & Foley,2007). Furthermore, it could be a more practical approach thanintensity scaling from the standpoint of consumer-led productdevelopment. Since CATA responses are directly linked to consum-ers’ perception of product characteristics, these responses could beutilized as supplemental data to maximize acceptance of the tar-geted products by consumers. CATA provides information onwhich attributes are detectable according to consumers and howthat may relate to their overall liking and acceptance. Understand-ing sensory characteristics in the process of new product develop-ment is of great importance, as failure to obtain correctinformation about the sensory attributes may lead to fast disap-pearance of the new products from the marketplace (Stone & Sidel,2007).

To understand the relationship between consumer and sensorydata, preference mapping is a useful method. Preference mappingis a widely used group of multivariate statistical techniques de-signed to optimize products by understanding the structure be-tween consumer preference and sensory data to identify driversof liking (Faye et al., 2006; Greenhoff & MacFie, 1999). Amongthe various product optimization mapping methods, the EuclidianDistance Ideal Point Mapping (EDIPM), an extension of Multidi-mensional Preference Mapping (MDPREF), is a new approach basedon a density analysis of individual consumer ideal product place-ments in the product configuration space (Meullenet, Xiong, &Findlay, 2007). In this approach, the ideal point of individual con-sumers is the point which minimizes the correlation betweenEuclidian distances to the products and hedonic scores.

Another optimization mapping technique, the Response SurfaceModel (RSM), proposed by Danzart, is based on external preferencemapping (Danzart, Sieffermann, & Delarue, 2004). Multidimen-sional representation of sensory stimuli is first created by sensory(i.e. external) data. The consumer data for individual consumers isthen regressed against the product coordinates in the sensoryspace to determine ideal points for both the individuals and thegroup (Meullenet, Lovely, Threlfall, Morris, & Striegler, 2008).

To investigate the efficacy of CATA scales within the sensoryenvironment, this study used ice cream as the testing medium.Ice cream is one of the most popular frozen desserts in the UnitedStates. The US ice cream market continues to grow and is expectedto be valued at over $10 billion by 2012 (Datamonitor, 2007). Va-nilla is the most popular ice cream flavor in the US and accountsfor almost half of all ice cream sales (Bodyfelt, Tobias, & Trout,1988). There are many companies producing and commercializingice cream in the US. In order to compete in this highly competitivemarket, it is crucial for ice cream manufacturers to understand thestrong and weak points of their products, and how consumers’ atti-tudes and preference patterns affect their products.

Table 1A list of 10 commercial vanilla ice cream products.

Brand Code Name/description Fat content (%)

Blue Bell A Homemade vanilla 13Blue Bunny B Premium all natural vanilla 10Ben and Jerry’s C Vanilla 24Best Choice D Vanilla 11Breyers E Natural vanilla 12Edy’s ‘‘Grand” F Rich and creamy vanilla 5Great Value G Vanilla 11Guilt Free H Vanilla 4Haagen-Dazs I Vanilla 28Yarnell’s J Homemade vanilla 15

The objectives of this study were to (1) assess the use of CATAattribute responses for 10 commercial vanilla ice creams as analternative to consumer attribute intensity ratings, and (2) com-pare CATA-generated preference maps to classical external mapsgenerated from traditional sensory profiles.

2. Materials and methods

2.1. Samples and sample preparation

Fifteen commercially-available ice creams were initially se-lected from local supermarkets for testing. Preliminary screeningof texture and flavor attributes eliminated five samples due tobrand replication and fat content, a popular indicator of ice creamquality, and the use of natural or artificial vanilla flavor. Ice creamswere selected so that various combinations of these quality factorswere represented in the study. The 10 remaining products, consist-ing of two high-fat products, six products with moderate fat con-tent and two low-fat products, are detailed in Table 1.

One scoop of each product was placed individually into a liddedwhite plastic container (45 mm diameter) coded with a three-digitrandom number. Samples were stored in a commercial-grade free-zer (TS-49, True Manufacturing Co., St. Louis, MO, USA) at 18 �C forat least 24 h prior to testing to ensure sample consistency. All sam-ples were tempered for 2 min at room temperature prior to servingfor both descriptive analysis and consumer testing. The 2 minincrement was determined to be the most appropriate temperingtime by observing the condition of the ice cream as a function oftime at room temperature. Samples were presented in a sequentialmonadic order to panelists according to a complete randomizedblock design, and the serving temperature (�12 + 2 �C) was strictlymonitored to maintain consistency (Bower & Baxter, 2003; Li, Mar-shall, Heymann, & Fernando, 1997).

2.2. Descriptive analysis

The 10 vanilla ice creams were evaluated for taste, aromatic, fla-vor, and texture attributes by a descriptive panel of 17 individualstrained by the Spectrum� method (Sensory Spectrum Inc., Chat-ham, NJ, USA). Panelists have over 100 h of training and an averageof 1000 h of testing experience. Two orientation sessions were con-ducted to familiarize the panelists with the samples. Flavor andtexture lexicons were developed in four sessions, as described inTables 2 and 3, respectively. The lexicons consisted of 23 total attri-butes specific to vanilla ice cream and definitions of each sensoryattribute with associated references. Panelists quantified all attri-butes on a line scale from 0 to 15 (Meilgaard et al., 2007). Unsaltedcrackers and water were provided for panelists to clean and rinsetheir palate between samples, and a 10 min break helped preventfatigue. The flavor attribute testing for all 10 products was

Flavor Manufacturer

Natural and artificial Blue Bell CreameriesNatural and artificial Wells’ Dairy, Inc.Natural Ben and Jerry’s Homemade Holdings, Inc. (Unilever)Artificial Wal-Mart Stores Inc.Natural UnileverNatural NestleArtificial Wal-Mart Stores Inc.Natural and artificial Yarnell Ice Cream Co.Natural NestleNatural and artificial Yarnell Ice Cream Co.

Table 2Flavor lexicon for vanilla ice cream.

Term Definition Reference Intensity

Basic tasteSweet The basic taste, perceived on the tongue, stimulated by sugars and

high potency sweetenersSolutions of sucrose inspring water

2–2.0% 5–5.0%

10–10.0% 16–15.0%Salt The basic taste, perceived on the tongue, stimulated by sodium

salt, especially sodium chlorideSolutions of sodiumchloride in spring water

0.2–2.0% 0.35–5.0%

0.5–8.5% 0.55–10.0%0.7–15.0%

Sour The basic taste, perceived on the tongue, stimulated by acids, suchas citric acid

Solutions of citric acid inspring water

0.05–2.0% 0.08–5.0%

0.15–10.0% 0.20–15.0%Bitter The basic taste, perceived on the tongue, stimulated by substances

such as quinine, caffeine, and certain other alkaloidsSolutions of caffeine inspring water

0.05–2.0% 0.08–5.0%

0.15–10.0% 0.20–15.0%

AromaticsVanillin The sweet, vanilla-like aromatic characteristic of ethyl vanillin or

imitation vanillasCooked milk The aromatic associated with the flavor of milk, heated to the

scalding pointUniversal scalea

Milky The aromatic associated with skim or whole milk products or milkderived products

Buttery/fat The aromatic associated with fresh butterfat; sweet cream Heavy whippingCream

Whole milk1/2 and 1/2

4.06.5

Cream cheese Heavy whip cream 9.0

FlavorNon fat dry milk The aromatic associated with boxed, nonfat dry milk or milk

reconstituted from dry milk solids, cardboardyCaramelized A sweet aromatic characteristic of browned sugars and other

carbohydratesOxidized The aromatics associated with slightly oxidized fats and oils Universal scalea

Woody/stick The aromatic associated with dry fresh cut wood; balsamic orbark-like

Metallic (1) The aromatic associated with metals, tinny or iron (2) a flatfeeling factor stimulated on the tongue by metal

Feeling factorsAstringent The feeling factor on the tongue or other skin surfaces of the

mouth described as puckering or drying(1) Alum solution (0.01%)(2) tea in 1000 ml springwater for 5 min

Alum –6.0 Tea –9.0

a Soda note (soda cracker 3.0), cooked apple note (applesauce 7.0), orange hit (orange juice 10.0), cooked grape note (grape juice 14.0), cinnamon hit (chewing gum 16.0).

396 L. Dooley et al. / Food Quality and Preference 21 (2010) 394–401

conducted on two consecutive days, followed by two consecutivedays for texture attribute testing.

2.3. Consumer testing

Eighty consumers were recruited to participate in the vanilla icecream test at the University of Arkansas Sensory Service Center(Fayetteville, AR). Qualification criteria included adults over18 years of age and vanilla ice cream product consumption at leastone to two times per week. The test was sequential monadic andsample presentation orders were balanced in each serving positionusing the Williams Latin Square design (Williams, 1949). Each con-sumer tested five samples the first day and tested the remainingfive the second day. The sessions were divided into 30 min incre-ments, with most consumers finishing within 20 min.

Consumers were asked to evaluate overall liking, appearance,flavor and texture attributes of each sample on a 9-point hedonicscale (1 = ‘‘dislike extremely”, 9 = ‘‘like extremely”). The attributesof scoopability, color, sweetness, vanilla flavor, creamy flavor,smoothness, melting in the mouth, melting in the bowl and hard-ness of each product were evaluated using the 5-point ‘‘Just AboutRight” (JAR) scale (data not shown or used in this study). Thoughnot used in these analyses, since this JAR data was collected, theauthors felt it necessary to mention. The final question of the sur-vey listed common ice cream attributes and asked the consumersto check all attributes that applied to the given sample (Table 4).

The CATA counts were totaled for each product (Table 5) and theresulting contingency table was used in subsequent analyses.

2.4. Statistical analyses

External preference mapping was performed according to Dan-zart et al. (2004) using either the descriptive sensory profiles orCATA counts to determine a group ideal point (Meullenet et al.,2008). First, principal component analysis was performed on meansensory profiles while correspondence analysis was performed onthe CATA counts. To determine the area of the map maximizingthe number of consumers satisfied, a quadratic model was con-structed (i.e. regressing hedonic scores against the principal com-ponent scores), and the area of acceptability for each consumerwas identified (i.e. area of the map where the hedonic score waspredicted above the mean score for each consumer). The area ofmaximum density was regarded as the ideal point solution for thismethod. Both analyses are considered here as external because theproduct configurations were obtained from data other than liking.

Since the descriptive profiles and CATA counts were scaled dif-ferently, the two data sets were standardized across all productsprior to external preference mapping to minimize differencesinherent to the scaling. To standardize the data, the data matrixhad products in columns and attributes in rows. The data was stan-dardized by columns so that the mean was zero and the variancewas one.

Table 3Texture lexicon for vanilla ice cream.

Term Definition Reference Intensity

Scoopability/manual firmness

The force required to cut the sample with yourspoon. 2 and 4 min

Evaluate the force required to removeone/spoonful of sample from the cup(soft–hard)

Peanut butter 4.0

Cream cheese 8.0Metal spoon sample andreferences

Hardness-oral The force required to compress the samplebetween the tongue and mouth roof. 2 min

Compress through sample one time withtongue. Evaluate the force required tofully compress the sample (soft–hard)

Peanut butter 4.0

Cream cheesePlastic spoon

9.0

Denseness The amount of air or fluffiness perceived in thesample. 2 min

Compress through sample one time withtongue. Evaluate the amount of airperceived in the sample (airy–compact)

Marshmallow cream 2.5

NougatPlastic spoon

4.0

Degree of ice The amount of ice crystals felt in the mouthduring the chew. 2 min

Compress 1/2 tsp of the sample. Evaluatethe amount of ice crystals perceived in thesample (no ice–much ice)

SherbetPlastic spoon

3.0

Smoothness The amount of particles perceived in thesample during the chew. 2 min

Manipulate the sample three times withtongue. Evaluate the amount of particlesperceived in the sample (not smooth–smooth)

Peanut butter 5.0

Cool whipPlastic spoon

14.0

Rate of melt The rate in which the ice cream changes formsfrom a solid to a liquid. 2 min

Place 1/2 tsp of sample in mouth andevaluate the rate in which the samplemelts (slow–fast)

Soy ice cream 5.0

Blue BunnyPlastic spoon

9.0

Mouth coat The amount and degree of residue felt by thetongue when moved over the surface of themouth. 2 min

Expectorate the sample and feel thesurface of the mouth with the tongue toevaluate (none–much)

Saltine 0.5

Ritz 2.5Pringles 5.0Mayo 9.0

Elasticity The degree to which the samples appears tohave an elastic/doughy impression. 4 min only

Stick spoon in the sample and pull out.After three pulls evaluate the amount ofsample pulled up by the spoon (none–much)

FrostingThree pullsPlastic Spoon

10.0

Table 4An example of check-all-that-apply (CATA) question.

Check all attributes that describe this sample:h Butteryh Sweeth Milk/dairy flavorh Custard/eggy flavorh Corn syruph Artificial vanillah Natural vanillah Creamy flavorh Softh Hardh Gummyh Icyh Creamy/smooth

L. Dooley et al. / Food Quality and Preference 21 (2010) 394–401 397

Internal preference mapping was also conducted as a point ofcomparison for descriptive analysis and CATA-based external pref-erence mapping. The comparison seems appropriate since theCATA-based external preference can be considered as a hybridmethod. Euclidian Distance Ideal Point Mapping (EDIPM, Meulle-net et al., 2008) was used for internal preference mapping. Theproduct configuration in the space was derived from principalcomponent analysis of the centered overall liking data.

Multiple factor analysis (MFA) was conducted using FactoMineRin R (v.2.6.2, 2008) to examine the similarities first between thedescriptive profiles and CATA counts (Figs. 1 and 2) and second be-tween the multivariate product configurations obtained from thethree preference mapping techniques employed. For each prefer-ence mapping technique, the coordinates of the ideal points inthe maps were estimated and used in the MFA analysis as illustra-tive data. MFA is a useful statistical technique to analyze the sim-ilarity of a set of observations explained by several different groupsof variables on comparable or contradictory scales (Abdi & Valen-tin, 2007). MFA is able to balance the influence of each variable,can compare multiple data sets, and can demonstrate patterns ofattribute correlation (Lê, Pagês, & Husson, 2008; Morand & Pagès,2005; Nestrud & Lawless, 2008).

3. Results and discussion

3.1. Comparison of product descriptions by CATA and descriptiveanalysis

Individual product maps were created by MFA using descriptivesensory profiles and consumer CATA counts (Fig. 1). Overall, MFAcomparing the characterization of the products by both profilesshowed agreement between the two methods, although only 51%

Fig. 1. Multiple factor analysis individual product plots using descriptive sensoryprofiles (D) and CATA counts (C). Product codes are listed in Table 1.

Fig. 2. Multiple factor analysis variable correlation circle obtained using descriptiveanalysis and CATA terms. Refer to Tables 2–4 for complete attribute descriptions.‘CS_c’ and ‘Cflavor_c’ represent creamy/smooth and creamy flavor, respectively, inCATA terms.

Table 5Total counts of check-all-that-apply attributes for each product.

Brand Soft Hard Gummy Icy Creamy/smooth

Buttery Sweet Milk/dairyflavor

Custard/eggyflavor

Cornsyrup

Naturalvanilla

Artificialvanilla

Creamyflavor

Blue Bell 42 20 5 37 31 44 58 60 45 10 25 39 45Blue Bunny 37 20 7 16 45 46 53 58 35 13 28 33 51Ben and Jerry’s 21 31 7 22 34 29 53 50 28 16 27 29 44Best Choice 29 29 4 28 37 43 61 62 32 13 34 33 54Breyers 8 43 7 61 10 17 57 52 25 11 37 37 24Edy’s ‘‘Grand” 25 38 4 20 35 39 61 59 26 16 28 43 53Great Value 51 11 12 8 59 38 46 58 21 12 27 31 52Guilt Free 34 22 3 29 35 30 52 57 22 23 19 45 35Haagen-Dazs 17 32 7 14 40 30 61 52 35 21 19 49 43Yarnell’s 43 16 4 14 51 21 60 54 22 15 24 42 47

398 L. Dooley et al. / Food Quality and Preference 21 (2010) 394–401

of the variability was explained by the first two MFA dimensions. Aand J showed the largest variance between the two methods, lar-gely due to the disagreement between the descriptive and con-sumer (CATA) maps for these two products (not shown). Asdiscussed by Pagès (2004), based on absolute scores of partial lineson each product for the first dimension, Fig. 1 indicates that prod-ucts B, H, I, and J were more characterized for the first dimensionby descriptive analysis profiles than CATA counts. If a line is drawnfrom the end of the descriptive and CATA vector for each of theseproducts to the x-axis, descriptive analysis rates higher on the firstdimension. Products E and G were more characterized by CATAcounts for the first dimension.

Fig. 2 shows the variable correlation circle obtained by MFAcomparing descriptive analysis profiles and CATA counts. The vec-tors showed a strong correlation for descriptive (d) mouth coat,smoothness and rate of melt, with CATA counts (c) perceived soft,creamy flavor and creamy/smooth, between degree of ice (d) andicy (c), elasticity (d) and gummy (c), and between caramelized(d) and corn syrup (c). The opposite vector directions for someCATA and descriptive descriptors with opposite meanings alsoshow agreement between the two methods. For example, hardness(c) was opposite to softness (c), and artificial vanilla opposite to

natural vanilla flavor. No correlations were observed between hard(c) and icy (c), and gummy (c), or between milky (d) and butterfat(d). Descriptive sensory profiles did not show any correlation be-tween bitter (d) and sweet (d) taste in the vanilla ice cream prod-ucts, while bitter (d) had negative correlation with salt (d) despiteits low loading. The most influential attributes (i.e. highest load-ings) were found to be sweet, bitter, vanillin, degree of ice, elastic-ity, and smoothness for descriptive sensory profiles. Icy, naturalvanilla, creamy/smooth, creamy flavor and artificial vanilla attri-butes for CATA consumer profiles played a relatively more impor-tant role in determining product locations in the map.

3.2. Preference mapping results and group ideal point locations

Each of the three preference mapping techniques employed(external mapping on descriptive data and CATA and internal pref-erence mapping) allow the identification of a group ideal productlocation in the maps. This point, in all three cases, is the locationin the map maximizing the percentage of consumers who wouldbe satisfied by a product placed at that location. Ideally, the threemethods would give approximately the same answer and this iswhat we seek to assess here.

Fig. 3. Results of external preference mapping using (a) descriptive analysis profiles (DD) and (b) CATA term counts (DC). The virtual product labeled ‘‘Opt” represents thelocation in the map maximizing the percentage of consumers satisfied (Danzart et al., 2004). Other product codes are as in Table 1. Values in parenthesis represent meanoverall liking values.

Fig. 4. Results of internal preference mapping. The virtual product labeled ‘‘Opt” represents the location in the map maximizing the percentage of consumers satisfied(Euclidian Distance Ideal Point Mapping (EDIPM), Meullenet et al., 2007). Other product codes are listed in Table 1. Values in parenthesis represent mean overall liking values.

L. Dooley et al. / Food Quality and Preference 21 (2010) 394–401 399

400 L. Dooley et al. / Food Quality and Preference 21 (2010) 394–401

The results of external preference mapping using descriptivesensory data (DD) and CATA counts (DC), and internal preferencemapping are graphically shown in Figs. 3 and 4. The first twodimensions of the maps explained 50.2% and 59.0% of the variancein descriptive profiles and CATA counts, respectively. The averageindividual consumer fit was similar for the DD map (R2 = 0.59)and the DC map (R2 = 0.61) (data not shown).

Overall, the products’ spatial representation on the first twodimensions seemed to differ. The product coordinates used to re-gress individual overall liking scores for the external maps wereobtained from two different data sources and the disagreement isnot too surprising. For the DD map (Fig. 3a), the ideal productwas closest to product F (OL = 6.7), followed by product D(OL = 6.8). Product E (OL = 5.5), least liked by consumers, was fur-thest from the other products, including the optimal product. TheDC map (Fig. 3b) placed the group ideal point near products A(OL = 6.3) and D (OL = 6.8), while product E was located on thefar left of the map by itself, showing agreement with the DDmap. Product D, which was most liked by consumers, was in clos-est proximity with the group ideal point for both maps. For internalpreference mapping, EDIPM placed the ideal point closest to prod-ucts A and D and furthest from product E (Fig. 4). This finding wassimilar to that of the DC map.

To more finely compare the level of agreement/disagreementbetween the three maps, MFA was employed using the first twodimensions of the three maps created. Fig. 5 represents the loca-tions of the 10 commercial vanilla ice cream products and the idealproduct (i.e. individual factor map) determined for each of thethree types of preference mapping on the first two MFA factors.

As shown in Fig. 5, the optimal product was in close proximitywith products A (OL = 6.3), D (OL = 6.8), and F (OL = 6.7), while E(OL = 5.5) was furthest from the remaining products and the idealproduct. Overall, MFA of the three product maps showed fairagreement between the approaches employed. However, the DDmap was contrary to the other two methods for products A and J,while the internal map showed dissimilarities to DD and DC, par-

Fig. 5. Multiple factor analysis individual product plots of the product configura-tions (first two dimensions) determined for external preference mapping usingdescriptive analysis data (DD) and CATA counts (DC) and internal preferencemapping (EDIPM). The virtual product labeled ‘‘Opt” represents the optimal productderived from the three mapping methods employed. Other product codes are listedin Table 1.

ticularly for product F. This could be explained by the fact thatproduct F was near the origin in both DD and DC maps, but was lo-cated in the far right upper quadrant in the EDIPM map (see Figs. 3and 4). One key result of this analysis is that optimal product loca-tion showed little variation between the three maps.

3.3. Ideal product profiles

Although the location of the group ideals in the three preferencemaps was reported to be fairly invariable, there is a need to deter-mine the sensory characteristics that should be exhibited by thegroup ideal products. If the three methods are in agreement, wewould expect the sensory profiles of the group ideals to be fairlysimilar. To determine the level of agreement between the prefer-ence mapping methods employed to determine the group idealsin the three maps, the CATA counts were regressed against the firsttwo dimensions of the product spaces created using descriptiveanalysis data, CATA data and internal preference mapping. Fromthese models, the ideal product CATA profiles were predicted. Thisis also known as reverse regression. Fig. 6a gives the fit of the CATAattributes for the three mapping methods employed. Overall, theCATA attributes were better fitted in the CATA space, particularlynatural vanilla, creamy flavor, creamy/smooth, soft and hard attri-butes (R2 > 0.6).

This is not surprising since this product space was derived fromthe CATA data. Overall, the CATA attributes were not as well fitted

Fig. 6. Ideal vanilla ice cream profiles according to descriptive, CATA and EDIPMdata using (a) CATA attribute fit (R2) and (b) normalized ideal CATA counts.

L. Dooley et al. / Food Quality and Preference 21 (2010) 394–401 401

in the product/preference space derived from internal preferencemapping while the quality of the fit was somewhere in betweenfor the product space created from descriptive analysis data. Thisresult is also not surprising since both the CATA and descriptiveanalysis product spaces are based on product sensory propertieswhile the internal preference mapping product space is based onliking data.

Fig. 6b represents the ideal sensory profiles in terms of normal-ized CATA counts for the three mapping methods employed. Thenormalization of these ideal profiles are easier to interpret thanideal raw CATA counts because they show if the ideal CATA countis above or below the mean of the observed counts. With normal-ized CATA counts, a positive value indicates that the ideal productwas above the mean CATA count across products for a given attri-bute. These profiles allow the determination of the level of agree-ment between the three mapping methods employed. From Fig. 6b,we conclude that the three mapping methods provided similaroptimal values for gummy, hard, soft, sweet, buttery, creamy/smooth, icy, creamy flavor and custard/eggy flavor. However, therewas disagreement for natural vanilla, artificial vanilla and corn syr-up. For these attributes, the CATA methodology recommended highnatural vanilla, low artificial vanilla and low corn syrup CATAcounts.

4. Conclusions

Overall, the characterization of the 10 commercial vanilla icecream products shows good agreement between descriptive sen-sory profiles and consumer-perceived CATA profiles. Moreover,the CATA attribute data applied to preference mapping gave simi-lar results to external preference mapping. The advantage of thistechnique is that the task asked of consumers is simple (i.e. whencompared to intensity ratings), and that the responses may bemore spontaneous than when intensities are rated. The limitationof this approach is that the optimal profile derived from the CATAmaps is in terms of response counts and not intensities as given bya trained panel or consumers using attribute intensity scaling.Since it can be hypothesized that the three maps created couldhave been more similar for product configurations, including opti-mal points, with additional dimensions being considered, the useof additional PCs may be of interest in future studies. AlthoughCATA questions seem to have some validity to characterize thesensory properties of products as perceived by consumers, manyquestions remain unanswered. Further studies may include theassessment of the effects of order and number of terms CATA ques-tions on attribute selection.

References

Abdi, H., & Valentin, D. (2007). Multiple factor analysis (MFA). In N. Salkind (Ed.),Encyclopedia of measurement and statistics. Thousand Oaks, CA: SagePublications.

Benedito, J., Cárcel, J. A., & Mulet, A. (2001). Cheese hardness by experts anduntrained judges. Journal of Sensory Studies, 16, 277–285.

Bodyfelt, F. W., Tobias, J., & Trout, G. M. (1988). Sensory evaluation of ice cream andrelated products. In F. W. Bodyfelt, J. Tobias, & G. M. Trout (Eds.), The sensoryevaluation of dairy products (pp. 166–226). New York: Van Nostrand Reinhold.

Bower, J. A., & Baxter, I. A. (2003). Sensory properties and consumer perception of‘home-made’ and commercial dairy ice cream. Journal of Sensory Studies, 18,217–234.

Danzart, M., Sieffermann, J.-M., & Delarue, J. (2004). New developments inpreference mapping techniques: Finding out a consumer optimal product, itssensory profile and the key sensory attributes. Seventh sensometrics meeting,Davis, CA, USA.

Datamonitor (2007). Ice cream in the United States: Industry profile. Available fromwww.datamonitor.com.

Deliza, R., Macfie, H., & Hedderley, D. (2005). The consumer perception of passion-fruit juice using free choice profiling. Journal of Sensory Studies, 20, 17–27.

Faye, P., Brémaud, D., Teillet, E., Courcoux, P., Giboreau, A., & Nicod, H. (2006). Analternative to external preference mapping based on consumer perceptivemapping. Food Quality and Preference, 17, 604–614.

Gonzáles-Tomás, L., & Costell, E. (2006). Sensory evaluation of vanilla-dairy dessertsby repertory grid method and free choice profile. Journal of Sensory Studies, 21,20–33.

González Viñas, M. A., Garrido, N., & Wittig de Penna, E. (2001). Free choice profilingof chilean goat cheese. Journal of Sensory Studies, 16, 239–248.

Greenhoff, K., & MacFie, H. J. H. (1999). Preference mapping in practice. In H. J. H.MacFie & D. M. H. Thomson (Eds.), Measurement of food preferences(pp. 137–166). New York: Aspen Publishers, Inc.

Guerrero, L., Gou, P., & Arnau, J. (1997). Descriptive analysis of toasted almonds: Acomparison between expert and semi-trained assessors. Journal of SensoryStudies, 12, 39–54.

Husson, F., & Pagés, J. (2003). Comparison of sensory profiles done by trained anduntrained juries: Methodology and results. Journal of Sensory Studies, 18,453–464.

Lancaster, B., & Foley, M. (2007). Determining statistical significance for choose-all-that-apply question responses. Seventh pangborn sensory science symposium,Minneapolis, USA.

Lê, S., Pagês, J., & Husson, F. (2008). Methodology for the comparison of sensoryprofiles provided by several panels: Application to a cross-cultural study. FoodQuality and Preference, 19, 179–184.

Lelievre, M., Chollet, S., Abdi, H., & Valentin, D. (2008). What is the validity of thesorting task for describing beers? A study using trained and untrained assessors.Food Quality and Preference, 19, 697–703.

Li, Z., Marshall, R., Heymann, H., & Fernando, L. (1997). Effect of milk fat content onflavor perception of vanilla ice cream. Journal of Dairy Science, 80, 3133–3141.

Meilgaard, M., Civille, G. V., & Carr, B. T. (2007). Sensory evaluation, techniques (4thed.). New York: CRC Press Inc.

Meullenet, J.-F., Lovely, C., Threlfall, R., Morris, J. R., & Striegler, R. K. (2008). An idealpoint density plot method for determining an optimal sensory profile forMuscadine grape juice. Food Quality and Preference, 19, 210–219.

Meullenet, J.-F., Xiong, R., & Findlay, C. J. (2007). Multivariate: And probabilisticanalyses of sensory science problems. Iowa, USA: IFT Press/Blackwell Publishing.

Morand, E., & Pagès, J. (2005). Procrustes multiple factor analysis to analyze theoverall perception of food products. Food Quality and Preference, 17, 36–42.

Nestrud, M. A., & Lawless, H. T. (2008). Perceptual mapping of citrus juices usingprojective mapping and profiling data from culinary professionals andconsumers. Food Quality and Preference, 19, 431–438.

Pagès, J. (2004). Multiple factor analysis: Main features and application to sensorydata. Revista Colombiana de Estadística, 27(1), 1–26.

Punter, P. H. (2008). Bridging the gap between R&D and marketing: The ideal profilemethod. In Third European conference on sensory and consumer research,Hamburg, Germany.

Seo, H.-S., Lee, S.-Y., & Hwang, I. (2009). Development of a sensory attribute pool ofbrewed coffee. Journal of Sensory Studies, 24, 111–132.

Stone, H., & Sidel, J. L. (2007). Sensory research and consumer-led food productdevelopment. In H. J. H. MacFie (Ed.), Consumer-led food product development(pp. 307–320). Cambridge, England: Woodhead Publishing Limited.

Williams, A. A., & Langron, S. P. (1984). The use of free choice profiling for evaluationof commercial ports. Journal of the Science of Food and Agriculture, 35, 558–568.

Williams, E. J. (1949). Experimental designs balanced for the estimation of residualeffects of treatments. Australian Journal of Scientific Research, Ser, A2, 149–168.

Worch, T., Lê, S., & Punter, P. H. (2008). How reliable are consumers? Comparison ofsensory profiles from consumers and experts. Ninth sensometrics meeting, St.Catherines, Ontario, Canada.