Consumer freshness perceptions of breads, biscuits ...

0 downloads 0 Views 290KB Size Report
Feb 27, 2009 - short biscuits to be fresher than pugliese and ciabatta bread, whilst ... biscuits or pugliese and ciabatta breads to be most fresh also per-.
Food Quality and Preference 20 (2009) 380–390

Contents lists available at ScienceDirect

Food Quality and Preference journal homepage: www.elsevier.com/locate/foodqual

Consumer freshness perceptions of breads, biscuits and cakes Samuel P. Heenan a,*, Nazimah Hamid a, Jean-Pierre Dufour a, Winna Harvey c, Conor M. Delahunty a,b a b c

Sensory Science Research Centre, Department of Food Science, University of Otago, Dunedin, New Zealand Food Science Australia, P.O. Box 52, North Ryde, NSW 1670, Australia New Zealand Institute for Crop and Food Research Limited, Christchurch, New Zealand

a r t i c l e

i n f o

Article history: Received 28 March 2008 Received in revised form 12 January 2009 Accepted 19 February 2009 Available online 27 February 2009 Keywords: Freshness Baked products Consumer vocabulary Descriptive analysis

a b s t r a c t Freshness is a holistic attribute of a food product, with meaning that most often includes how recently produced or harvested a food currently is, to what extent it has been preserved, and it is the opposite of stale or spoiled. The sensory properties of foods contribute very significantly to consumer freshness perceptions, but these properties are not easily identified and described as they will vary considerably between different product types. The aim of this investigation was to improve the understanding of the sensory properties that contribute to the perceived freshness of different baked product types, including breads, biscuits and cakes. Two separate consumer studies were conducted. In the first study, consumers evaluated product freshness based on appearance, flavour and texture and in the second, consumers only evaluated product odour. Qualitative information detailing the product characteristics that consumers most associated with freshness was collected following assessment at each study. In order to better understand the sensory properties associated with product freshness, a trained sensory panel carried out descriptive analysis with the baked products, using a defined vocabulary. Individual differences in perceived product freshness were represented on two-dimensional Freshness Maps generated using Principal Component Analysis. Results showed the relative similarities and differences between the product types on the basis of individual freshness perceptions. Consumers used similar terms to describe product freshness, but different terms were associated with specific product types. Relationships between consumer freshness ratings and descriptive analysis showed attributes that were positively and negatively associated with freshness for each product type when evaluated following either consumption, or by odour only. It was found that sensory characteristics perceived to be fresh in one product type was not necessarily fresh in another. Comparisons between consumer and sensory vocabularies showed specific appearance, odour, taste and textural cues that consumers associated with product freshness were used to differentiate the same products cognitively. Ó 2009 Elsevier Ltd. All rights reserved.

1. Introduction More often than not, the perceived freshness of baked products are considered as one of the key determinates of acceptance and choice (Nordmark, 1999; Pickford, 1999). Pervious research on purchase intention of baked products indicates that 80% of bakery sales are impulse purchases induced by perceived freshness (South, 1989). The sensory properties of baked products, in terms of appearance, odour, taste and texture, without doubt contribute strongly to consumer freshness perceptions, although it is not clear what sensory properties contribute to the freshness of different baked product types from a consumer’s perspective. However it is likely, that this appreciation of freshness is implicitly learned,

* Corresponding author. Tel.: +64 3 479 5463; fax: +64 3 479 7567. E-mail addresses: [email protected], [email protected] (S.P. Heenan). 0950-3293/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.foodqual.2009.02.008

whereby it is encoded and recalled without the subject’s awareness. Research concerning the implicit sensory memory suggests that non-conscious decisions relate highly to specific food stimuli (Degel & Koster, 1999; Degel, Piper, & Koster, 2001; Koster, Prescott, & Koster, 2004). Consumer perceptions of a product involve to a large extent what is remembered of the product itself, and not just what is consciously perceived when encountering that product. For example, individuals find it difficult to identify and describe odours and flavours but are more likely to recall a place and time these odours and flavours were experienced previously (Koster et al., 2004). Hence the freshness perception may not be easily described by consumers, particularly as it is likely to vary between different product types. According to Cardello and Schutz (2003), type of food influences freshness perceptions, while the meaning of freshness varies between individuals. A good understanding of food product freshness in relation to sensory properties has been demonstrated by asking consumers to rate freshness while evaluating

S.P. Heenan et al. / Food Quality and Preference 20 (2009) 380–390

food (Peneau, Hoehn, Roth, Escher, & Nuessli, 2006). In their investigation, consumers discriminated differences in the level of freshness for apple cultivars that were stored under different conditions. In addition, Fillion and Kilcast (2000) showed that the sensory properties of fruits and vegetables were important in distinguishing consumer perceptions of freshness. Multidimensional data analysis techniques can be applied to determine relationships between sensory characteristics defined and measured by a trained sensory panel and individual consumer perceptions, allowing the sensory characteristics associated with consumer perceptions of products to be identified (MacFie & Hedderley, 1993; Martens & Martens, 1986). This data analysis approach has been applied to understand consumers’ perception of bread (Hersleth et al., 2005), and cheese (Hersleth et al., 2005). In particular, consumer perceptions of freshness in apples (Peneau, Brockhoff, Hoehn, Escher, & Nuessli, 2007; Peneau et al., 2006); creaminess in dairy products (Richardson-Harman et al., 2000; Tournier, Martin, Guichard, Issanchou, & Sulmont-Rosse, 2007) and refreshing in gel model systems (Labbe, Gilbert, Antille, & Martin, 2009), have been studied using descriptive sensory analysis in parallel with consumer research, and by relating data using multivariate techniques. According to Carr, Craig-Petsinger, and Hadlich (2001) relationships between consumer descriptions of products and their sensory character provide a valuable insight into understanding how consumers perceive complex product image attributes such as wholesome, comforting and sophisticated. Moreover, Roininen, Arvola, and Lahteenmaki (2006) demonstrated that consumers’ perception of foods can be defined by word association techniques when combined with scored responses. This provides both a qualitative interpretation as well as a quantitative understanding of food perception. The aim of this investigation was to identify and compare the sensory properties associated with consumers’ perception of freshness in breads, cakes and biscuits. The baked products were evaluated in terms of appearance, ortho- and retro-nasal olfactory information as well as taste and tactile information (study 1), and product odour only, to include only ortho-nasal information (study 2). Secondly, important sensory qualities that consumers most associated with product freshness from each study were compared. In addition, a trained panel was used to identify and measure sensory differences between baked products in terms of appearance, odour, taste, flavour, texture and aftertaste. Finally, the relationship between consumer ratings of freshness and sensory characteristics defined and measured by descriptive analysis were investigated. 2. Materials and methods 2.1. Experimental plan Two separate consumer studies were performed. In both studies, participants were presented with 20 baked product types consisting of breads (n = 10), cakes (n = 5) and biscuits (n = 5), and were asked to evaluate product freshness. In both studies participants were balanced in terms of age and gender. In the first study, participants (n = 115) evaluated the freshness of baked products on the basis of appearance, odour, flavour and texture. In the second study, participants (n = 102) from which 82% (n = 84) consumers from the first study evaluated the freshness of baked products by odour only. All samples were selected based on differences between the major raw ingredient components of the dough (flour, water, yeast, salt, sugar, fat, eggs) as shown in Table 1. Variations within these ingredients ensured that the products differed in their sensory characteristics for odour, taste, flavour and texture. All breads were baked and packaged by local bakeries on the morning of testing, and were tested within 4 h from baking. All

381

cakes were baked and packaged by a commercial bakery on the day before testing, and were tested within approximately 16 h from baking. Biscuits were purchased from the same commercial bakery as cakes in packages and tested prior to their best before date. In terms of freshness, time from baking to consumption was chosen to represent what is typically experienced by consumers in New Zealand. Consumer testing was conducted in individual booths at the Sensory Science Research Centre, University of Otago between the hours of 8:00 am and 9:00 am. Testing took place on a blind basis. All products were assigned a three digit random code and were presented unbranded at room temperature. The samples of breads, cakes and biscuits were served to participants in four sets of five samples (one set for cakes, one set for biscuits, and two sets for breads). This was done to ensure product types were evaluated separately. To balance the effect of serving order and carry over, the three product types and five products within the four sets were balanced in accordance to the Williams Latin Square (MacFie & Bratchell 1989). In the first study, where products were consumed, breads and cakes were sliced into portions, consisting of both the crust and the crumb, while one whole biscuit was presented. In the second study, 25 g of each product was placed into 250 ml amber glass bottles to obscure the products visual appearance, and bottles were sealed with 38 mm screw caps. Each bottle was assigned a three digit random code. Recruitment of participants was achieved by advertising in local papers. Participants were recruited on the basis that they regularly consumed baked products (i.e. more than one time per week), had lived in New Zealand for a period of five years or longer, as well as had no known allergies or difficulties in perceiving smell, taste or swallowing foods. It has been reported that regular consumers of specific food types are familiar with what constitutes freshness based upon previous experience (Cardello & Schutz, 2003). In a separate study, a trained descriptive panel (9 females and 2 males aged between 24 and 55 years) carried out descriptive analysis with the product types; breads (n = 10), cakes (n = 5) and biscuits (n = 5). Analysis was carried out in duplicate using a defined vocabulary developed for each product type that described product appearance, odour, taste-flavour, texture and aftertaste (Stone, 2004). Each attribute was evaluated on a continuous unstructured 100 mm line scale anchored with low (1) to high intensity (100). 2.2. Consumer testing Testing was carried out over a period of six weeks. In the first three week block, participants rated freshness after consumption (study 1), and in the second three week block by odour alone (study 2). Prior to evaluating products in each study, a standard context was read aloud by the investigator, which questioned the authenticity of product ingredients in relation to product freshness. Numerous studies have found that consumers are highly influenced by product information in relation to their cognitive and affective reactions when evaluating food (Guinard et al., 2000; Johansson, Haglund, Berglund, Lea, & Risvik 1999; Martins, Pelchat, & Pliner, 1997; Tuorila, Meiselman, Bell, Cardello, & Johnson, 1994). For example, the liking of the perceived sensory quality of bread was influenced by a read context, which questioned the authenticity of bread ingredients and their associated health benefits (Kihberg, Johansson, Langsrud, & Risvik, 2005). As a consequence, consumers may be focused by the content of the information read to them, which in turn reduces the effects of non sensory differences (i.e. cognitive differences, social differences) so as to obtain a more uniform response between the participants’ perception of food (Frandsen, Dijksterhuis, Brockhoff, Nielsen, & Martens, 2003, 2007). The intention of the context used in the current study, was to focus all participants on the freshness concept and evoke an affective response.

382

S.P. Heenan et al. / Food Quality and Preference 20 (2009) 380–390

Table 1 Baked product ingredient composition. Baked product

Wheat flour (%)

Water (%)

Yeast (%)

Salt (%)

Fat (%)

Sugar (%)

Egg (%)

Bread Bagel Brioche Ciabatta Croissant Foccacia Multi grain Pugliese Rye Sourdough White bread

57.2 42.3 54.0 43.5 47.3 21.9 50.3 40.3 58.6 54.3

36.6 0 38.0 28.5 38.0 37.3 36.8 28.4 37.2 32.4

0.50 1.70 0.60 1.10 0.50 0.70 0.60 0.50 – 1.10

1.20 0.90 2.00 0.90 1.20 1.10 2.00 1.00 1.20 0.90

2.50 25.5 3.20 20.2 13.0 7.10 5.00 4.00 1.00 4.30

2.00 4.20 2.40 5.80 0 0 5.30 19.2 2.00 7.00

0 25.5 0 0 0 0 0 0 0 0

Cakes Angel Butter Madeira Semolina Wholemeal

15.4 27.0 37.4 13.2 18.6

0 0 0 0 0

0 0 0 0 0

0.30 0.50 0 0 0.50

16.7 29.6 22.2 12.3 8.40

41.6 32.1 26.1 24.0 23.0

26.2 10.8 14.6 16.0 13.5

Biscuits Bran Cream Malt Oat Short

18.4 62.3 41.9 27.8 46.8

4.00 13.0 3.00 4.00 5.00

0 0 0 0 0

1.70 3.00 1.20 1.70 2.10

21.9 8.34 13.5 17.7 28.9

17.2 4.00 19.4 1.90 22.1

16.0 0 16.0 0 0

To make sure all participants received the same information, the context was also displayed visually on a large screen using a computer projector. Due to the participants’ regular consumption of baked products, it was assumed they had a vested interest in what the product contained. From this aspect, the following context was constructed: ‘‘We have invited you here today to evaluate a selection of baked products. Your task is to judge the freshness of each product, and indicate your decision by using the line scale provided. We are interested in your freshness judgement of these products because some of them have been manufactured using ingredients that contain preservatives, stabilisers and emulsifiers, whereas the remainder of the products have been manufactured using the finest natural ingredients. The issue is whether products produced with ingredients that contain preservatives, stabilisers and emulsifiers are less fresh than products produced with natural ingredients. In order to find this out, we ask you to evaluate all baked products, and judge their degree of freshness”. The intention was to ‘‘bias” the consumers towards discriminating between fresher and less fresh products. In fact some products were manufactured using pre-mix ingredients, whereas others were not. In each study, participants were asked to judge sample freshness using a 150 mm line scale with log intervals (0.2 log series) to separate adjective labels from ‘‘not at all fresh” (0 mm), ‘‘Barely fresh” (10.5 mm), ‘‘Moderately fresh” (26 mm), ‘‘Very fresh” (52.5 mm), ‘‘Extremely fresh” (89 mm), and ‘‘Greatest freshness imaginable” (150 mm) (Forde and Delahunty, 2004). This scale is similar in principal to the labelled affective magnitude scale (LAM), which has demonstrated good reliability and discriminative sensitivity for assessing product liking (Schutz & Cardello, 2001), although in this case assessors did not scale the labels using magnitude estimation. This scale was chosen as it provided a fixed end point ‘‘greatest imaginable” that avoids end affects and commonly aligns consumers to a fixed rule. Retasting or smelling of products was allowed, although participants were advised that this should be limited to prevent becoming too satiated. Filtered water was provided and participants were instructed to rinse their palates between samples with a break

of 30 s. In addition, there was a forced 5-min interval between each set of five samples to reduce carryover effects. After evaluating all products, participants were asked to fill out a questionnaire that required them to recall the context read out at the beginning of the session. This was done to confirm if participants were aware of the context while assessing sample freshness (Frandsen et al., 2003, 2007). Consumers were then asked to record their frequency of baked product consumption per week. Finally consumers were asked an open ended question regarding what qualities that they most associated with freshness based on their experiences with each baked product type (i.e. breads, cakes and biscuits). This way of questioning (immediately after consumption) avoids potential bias caused by relying on long-term memory (Ragaert, Verbeke, Devlieghere, & Debevere, 2004). All assessments were conducted in individual booths at the Sensory Science Research Centre, University of Otago, between the hours of 8:00 am and 9:00 am. At the end of each study, as a token of appreciation, participants received a petrol or grocery voucher to the equivalent of $ 20. 2.3. Descriptive analysis A panel of 11 assessors (9 females and 2 males, aged between 24 and 55 years) evaluated baked products using descriptive analysis (Stone & Sidel, 2004). Panellist were selected, screened and recruited from the population of Dunedin, New Zealand, and trained according to the guidelines in ISO 8586 (1994). Consensus on the definition of terms for each product type was reached by assessors, over 16 2 h training sessions. For evaluation, approximately 30 g of each sample, in a 3-digit random coded glass covered with a glass lid were presented to assessors in individual booths under white light at room temperature. The sample types were evaluated on different days to ensure they were assessed separately. To balance the effects of serving order and carryover, duplicate samples were presented in accordance to Williams Latin Square (Macfie & Bratchell, 1989). To reduce the occurrence of carryover assessors were instructed to break for a 2 min interval between samples. In addition, each assessor was provided with filtered water and un-salted crackers and was asked to cleanse their palate between tastings. For each product type, assessors received a list of attributes, with definitions to aid them during product assessment.

S.P. Heenan et al. / Food Quality and Preference 20 (2009) 380–390

2.4. Consumer data analysis For all products, participant ratings on the labelled 150 mm line scale were measured geometrically to produce arbitrary freshness values (mm). In both studies, product freshness was evaluated separately by Univariate analysis of variance (ANOVA) using SPSS 12.0.1 (SPSS Inc., Chicago, USA). Participant freshness scores were taken as the dependant variable, with products as factors to be tested. From individual consumer ratings, mean freshness scores were taken for each product and compared using post-hoc Tukey honestly significant difference (HSD) testing (p < 0.05). To illustrate the intensity and direction of each participants’ freshness perception of products, and to discriminate between different product types based on individual freshness ratings, Principle Component Analysis (PCA) was performed on individual data from each study using the Unscrambler statistical software version 9.1 (CAMO, AS, N-7041 Trondheim, Norway). Two underlying dimensions were used to illustrate the differences between the products investigated. From each study, qualitative results from the questionnaire were recorded separately. The proportion of participants who recorded correct or near correct versions of the context read out prior to product evaluation were represented. Qualitative terms that participants’ most associated with freshness in breads, cakes and biscuits, were counted and shown as the most frequently used terms. Terms recorded by more than ten consumers were tabulated. In this case, terms used by less than ten participants were considered as idiosyncratic (Faye et al., 2006; Hersleth, Berggren, Westad, & Martens, 2005a). Words with limited meaning (i.e. the, for, and, of, to) were removed from the analysis (ten Kleij & Musters, 2003). No selection was done on the meaning of terms. To evaluate possible interactions with consumer demographic effects, freshness ratings from each study were analysed separately as a dependant variable in a three-way ANOVA model with age, gender and frequency of consumption (i.e. once a week and three or more times a week) as factors. For ANOVA models, significance was defined at p < 0.05. 2.5. Descriptive data analysis Duplicate descriptive sensory analysis scores for each attribute were averaged and analysed using one-way ANOVA and Fisher’s least significant difference (LSD) test. Attributes that did not significantly discriminate between the different breads, cakes and biscuit types were not used in further analysis. Further examination of the sensory characteristics that defined consumer perceptions of freshness for different baked product types was carried out using partial least squares regression type 1 (PLSR1; (Martens, 2001). Six different PLSR models were created to investigate relationships between sensory attributes (X-variables) and consumer freshness scores (Y-variables) for breads, cakes and biscuits following either consumption or odour evaluation only. Full cross validation was used to select the optimum number of PLS factors for predicting the Y-data sets. The calibration and validation coefficients that expressed model fit in X and Y, and ability to predict new data were monitored. For each model, the regression coefficients (b-coefficients) were graphed to determine the contribution of each X-variable in predicting Y-variables. Variables that contributed little and/or displayed high levels of uncertainty estimates were removed (Martens, Bredie, & Martens, 2000; Martens & Martens, 2000). Subsequent models were recalibrated and final PLSR models were selected based on the root mean square error of prediction (RMSEP). The RMSEP represents the average prediction error expected for new samples based on the original response variables (consumer freshness scores and sensory attributes on a scale from 0 to 150 mm). In addition, RMSEP

383

was used to study the relationship between modelling error and estimation error to determine the optimum number of principal components (PCs) to be used. To avoid dependence on the measured units, X-variables were standardised (1/standard deviation) prior to modelling (Martens & Martens, 1986).

3. Results and discussion 3.1. Consumer freshness evaluations Preliminary analysis of data collected in Study 1 using ANOVA showed that products differed very significantly (d.f. = 19, F = 34.47, p < 0.001) in their perceived freshness. By consuming and rating product freshness, participants were able to discriminate significant differences between breads (d.f. = 9, F = 13.11, p < 0.001), cakes (d.f. = 4, F = 95.34, p < 0.001) and biscuits (d.f. = 4, F = 6.13, p < 0.001). Post-hoc testing (Table 3) revealed the freshest breads to be foccacia, croissant, and rye, whilst breads perceived as least fresh were multigrain and sourdough. For cakes, consumers perceived angel cake and butter cake to be the most fresh and Madeira, wholemeal and semolina to be the least fresh. Short, oat, bran and cream biscuits were considered as most fresh, whilst malt biscuit was least fresh (Table 3). As the time from baking to consumption was near equal for all breads, all should be considered optimally fresh from a products perspective. Subsequently the differences observed in perceived freshness are most likely related to bread type and their sensory quality. In addition, all cakes were evaluated within approximately 16 h from baking and therefore perceived freshness differences are more than likely a result of cake type in relation to sensory quality. While biscuits were evaluated within their shelf-life, time from baking to consumption was not defined for each biscuit type. This was a limitation of the research, as reported differences between biscuits types may not fully reflect ‘‘perceptual” differences from a perceived freshness standpoint, but may rather reflect physical/ sensory changes that occurred in the biscuit types as a result of different times from baking. Nevertheless, the type of product, independent of production time, has been reported to influence judgements of freshness particularly as the sensory properties responsible for the perception of freshness are product specific (Cardello & Schutz, 2003). In light of this understanding, perceived freshness is subjective and baked products are deemed fresh, or at varying degrees of less fresh, according to consumer judgment. This measure can be independent of age, or time from baking of the baked product (Hebeda & Zobel, 1996). In order to illustrate differences between product types on the basis of individual consumer perceptions of freshness, PCA was carried out. The PCA, or ‘‘freshness map”, described 38% and 18% of the variation on principal components 1 (PC1) and 2 (PC2), respectively (Fig. 1). The freshness map illustrates important information on the intensity and direction of individual freshness judgements, which determine the position of all the baked products relative to each other (Greenhoff & MacFie, 1994). Distance and direction from the map origin to any consumer point indicates the direction of increasing freshness perception for that individual (Helgesen, Solheim, & Naes, 1997). Based on the distribution of individual freshness perceptions in this study, participants strongly agreed that angel cake, croissant, foccacia bread, rye bread and brioche located on the positive axis of PC 1, were most fresh, while semolina cake, wholemeal cake, madeira cake and malt biscuit located on the negative axis of PC 1 were less fresh (Fig. 1). PC 2 explained some differences between participants in what they perceived to be the freshest products (Fig. 1). In this case, participants located along the positive axis of PC2 perceived bran and short biscuits to be fresher than pugliese and ciabatta bread, whilst

384

S.P. Heenan et al. / Food Quality and Preference 20 (2009) 380–390

Principal Component 2 (18 %)

1.2 Bran-biscuit Short-biscuit Cream-biscuit Oat-biscuit

0.6

Angel-cake

Butter-cake Malt-biscuit Semolina-cake

0

Croissant-bread Brioche-bread

Wholemeal-cake Madeira-cake

Foccacia-bread Bagel-bread Rye-bread

-0.6

Multigrain-bread Sourdough-bread

White-bread Ciabatta-bread Pugliese-bread

-1.2 -1.2

-0.6

0

0.6

1.2

Principal Component 1 (38 %) Fig. 1. Principal component analysis (PCA) of baked products perceived freshness after consumption.

participants on the negative axis of PC2 perceived these breads to be fresher than the biscuits. Pugliese bread is described as traditional sourdough bread, produced in southern Italy (De Vuyst & Neysen, 2005). Nevertheless, product freshness differences along PC 2 showed far less explained variance than those separated along PC 1. In this instance, participants who scored either bran and short biscuits or pugliese and ciabatta breads to be most fresh also perceived angel cake, croissant, foccacia and rye bread as fresh. By evaluating product odour only in the second study, participants were also able to discriminate significant differences in product freshness (d.f. = 19, F = 14.45, p < 0.001). Separate ANOVA results showed that participants perceived significant differences between types of breads (d.f. = 9, F = 13.76, p < 0.001), cakes d.f. = 4, F = 15.10, p < 0.001) and biscuits (d.f. = 4, F = 13.48, p < 0.001). Croissant and brioche smelled the freshest, whilst bagel, white bread, and sourdough were considered as less fresh (Table 2). Participants distinguished the smell of madeira and butter cake as fresh and angel and semolina cake as less fresh. The smell of short and malt biscuits were perceived as most fresh, while oat, cream and bran biscuits were scored as less fresh. The freshness map for the second study explained 27% and 12% of variation on PC 1 and PC 2, respectively. In terms of freshness, participants

Table 2 Descriptive vocabulary and definitions used by trained assessors to evaluate baked products. Attribute

Product

Definition

Appearance Crumb darkness Crust darkness Crumbly Crust Smoothness Moisture Porosity Speckled

Biscuits, breads, cakes Biscuits, breads, cakes Biscuits Biscuits, breads Breads, cakes Breads, cakes Breads

Degree of colour darkness in the crumb ranging from white to dark brown Degree of colour darkness in the crust ranging from light brown to dark brown Amount of product particles that have separated from the crumb and crust The visual appearance of crust contour from smooth to rough The degree of surface wetness from dry to wet The extent of perforation of the product surface, this encompassing the holes, cracks allowing the permeation of air The amount of visual flecks through out the product

Odour Caramel Citrus Coconut Dairy Dusty Eggy Grain Malty Musty Nutty Vanilla Toasted Yeasty

Biscuits, Cakes Biscuits, Biscuits, Breads Biscuits, Biscuits, Biscuits, Breads Biscuits, Biscuits, Biscuits, Breads

Distinctive odour representative of heated sugar Primary aroma of lemon, citron, lime, orange, grapefruit, etc. Citrus fruit – a fruit of such tree. The aromatics sensation perceived when smelling shredded desiccated coconut Aromatics associated with products produced from cows milk The aromatics associated with standard baking flour, e.g., wheat flour Aroma typical of hard boiled eggs An aromatic impression of cereal derived products, typically rye, wheat, oats, cornmeal and barley Sweet aromatic typical of condensed milk, toffee, and/or malt The aromatics associated with wet grain and/or damp earth The aromatic characteristics of mixed nuts, e.g., walnuts, hazelnuts, brazil nuts and pine nuts Typical aromatics associated with Vanilla bean The odour impression of bread and crumb after baking/heating Odour associated with aromatic exchange from yeast fermentation

Taste/Flavour Bitter Buttery Doughy Oily Salty Seedy Sour Sweet

Breads Biscuits, breads, cakes Cakes Breads Breads Breads Breads Biscuits, breads, cakes

Fundamental taste sensation of which caffeine or quinine are typical The sweet aromatic flavour impression of products produced from buttery, e.g., puff pastry Unbaked flavour impression associated with moist wheat flour The overall flavour impression of oil Fundamental taste sensation elicited by sodium chloride The overall flavour impression elicited by plant derived seeds, e.g., sunflower Fundamental taste sensation evoked by acids, e.g., tartaric acid Fundamental taste sensation of which sucrose is typical

Oral-Texture Adhesiveness Coarse Crunchy Fat Hardness Moisture

Biscuits, Breads Biscuits Biscuits, Biscuits, Biscuits,

Force required to remove sample completely from the palate, using the tongue during consumption Degree to which particles reside on the palate during consumption Degree of brittleness when force is applied between the molars Degree of residual oiliness left on the oral cavity after swallowing the sample Force required to bite completely through sample placed between the molars Amount of moisture perceived on the surface of the product, when in contact with the oral cavity

After-Taste/flavour Bitter Burnt Sour Sweet

Breads, cakes Breads, cakes Breads Biscuits, cakes

cakes cakes breads cakes breads, cakes breads breads cakes breads, cakes

breads, cakes

breads, cakes breads, cakes breads, cakes

Residual taste sensation following consumption typical of and/or quinine are typical The sharp, acrid note of product that had been over cooked Residual taste sensation following consumption evoked by acid, e.g., tartaric acid Residual taste sensation following consumption evoked by sucrose

S.P. Heenan et al. / Food Quality and Preference 20 (2009) 380–390 Table 3 Mean* ratings of freshness on a 150 mm labelled scale of baked products assessed after consumption (n = 115) and smelling (n = 102). Baked Sample Bread Bagel Brioche Ciabatta Croissant Foccacia Multi grain Pugliese Rye Sourdough White

Cake Angel Butter Madeira Semolina Wholemeal

Biscuits Bran Cream Malt Oat Shortbread

Product by consumption

Product by odour

50.94b 57.04cd 54.23bcd 67.61de 69.12e 38.02a 55.86bc 65.17de 37.10a 53.89bc F = 13.11 p < 0.001

28.38a 54.14e 35.97bc 57.90e 34.09ab 40.65bcd 36.87bc 47.71ed 32.76ab 32.16ab F = 13.76 p < 0.001

71.06c 48.87b 18.81a 21.47a 19.55a F = 95.34 p < 0.001

23.16a 39.14b 42.80b 24.82a 29.51a F = 15.10 p < 0.001

47.98b 46.03b 34.23a 49.67b 53.23b F = 6.13 p < 0.001

32.02a 27.06a 41.55b 24.21a 40.69b F = 12.48 p < 0.001

*

Means with same letters are not significantly different (Tukey multiple comparisons test p < 0.05).

had a tendency to agree that brioche and croissant were most fresh along the positive axis of PC 1, whilst semolina cake, foccacia bread and angel cake were the least fresh (Fig. 2). Similar to the first study, participants’ freshness perception along PC 2 were separated by either ciabatta or multigrain breads along the positive axis or bran and short biscuits along the negative axis. It should be noted that commercial products evaluated in this study were presumed to be presented as fresh in relation to what is typically

Principal Component 2 (12 %)

1.2

Multigrain-bread Ciabatta-bread White-bread Sourdough-bread Rye-bread Pugliese-bread

0.6

Bagel-bread

0.0

Wholemeal-cake Semolina-cake Foccacia-bread Butter-cake Angel-cake

Maderia-cake Croissant-bread Brioche-bread Malt-biscuit

Oat-biscuit

-0.6 Cream-biscuit Bran-biscuit Short-biscuit

-1.2 -1.2

-0.6

0

0.6

1.2

Principal Component 1 (27 %) Fig. 2. Principal component analysis (PCA) of baked products perceived freshness after smelling only product odour.

385

available to consumers. Therefore the present study was not designed to evaluate product freshness in relation to the elapse of time from baking. Subsequently more work is required to investigate the perceived freshness of individual product types at different intervals from the time of baking. From evaluating products following consumption or odour only, freshness maps from both studies showed participants agreed that croissant, brioche, rye bread, butter cake and short biscuit were the most fresh. Observations of ingredient composition (Table 1) indicate that croissant, brioche, rye bread, butter cake and short biscuit contained either a high level of fat and/or sugar. It is widely known that changing the food matrix composition and manufacturing conditions of complex food systems such as breads, cakes and biscuits will influence the perception of the sensory characteristics (Abdallah, Chabert, Le Roux, & Louis-Sylvestre, 1998; Baardseth, Naes, & Vogt, 1995; Brauss, Balders, Linforth, Avison, & Taylor, 1999; Cardello & Schutz, 2003; Hebeda & Zobel, 1996; Heinio, Urala, Vainionpaa, Poutanen, & Tuorila, 1997). Abdallah et al. (1998) showed that the level of fat and sugar in biscuits and cakes enhanced flavour intensity, which in turn was very closely related to pleasantness perceptions. However there were some clear differences between product freshness scores in both studies. These differences were evident in foccacia bread and angel cake, which were rated significantly (p < 0.05) less fresh when only odour was evaluated. In contrast, multigrain bread, madeira cake and malt biscuit were perceived to be significantly (p < 0.05) more fresh when evaluated by odour only (Table 3). This difference might be expected, given participants’ inability to evaluate products based on appearance, taste, flavour and texture when evaluating the product based on odour only. Previous observations have shown that certain foods appear to be sensed differently, depending on the mode of presentation (Hummel et al., 2006). In their study attention was drawn to the interaction between the chemosensory modalities (i.e. taste, smell and trigeminal system), when food samples were evaluated separately by ortho and retronasal olfaction. Researchers have suggested that olfactory stimulation via the mouth and that via the nose can result in different perceptual experiences (Rozin, 1982). This suggestion is further supported by Rankin and Marks (2000) who reported that participants judged perceptual experiences of the olfactants, vanillin and orange in sucrose solution to be different qualitatively when evaluated in separate studies by tasting or smelling only. Variation in the freshness perception of foccacia bread, angel cake, multigrain bread, madeira cake and malt biscuit when evaluating products following consumption to smelling odour only suggests that the odour quality is in fact one component of the freshness perception. Notably, participants may have been influenced largely by the textural characteristics of products when judging freshness following consumption. Although attention was taken to recruiting the same consumers from the first study to carry out the second study, some variation between consumer responses might be expected. In addition, as consumer testing was conducted on different days with a different batch of products, some further variation among product sensory character and judgements of freshness might be expected. Previous studies have established that both appearance and textural attributes were important for judging fruit and vegetable freshness, whilst other characteristics of freshness such as odour, was more elusive (Fillion & Kilcast, 2000; Peneau et al., 2007). Under these circumstances, it can be assumed that the sensory sensations perceived by appearance, odour, taste, flavour, and oral texture interact to a give a single impression, which acts as the basis for a freshness decision. Nevertheless this investigation provided the opportunity for participants to focus on the odour properties of baked products, thereby providing a more detailed

386

S.P. Heenan et al. / Food Quality and Preference 20 (2009) 380–390

understanding of how odour contributed to baked product freshness. 3.2. Qualitative information from consumer freshness evaluations Information from the post experimental questionnaire indicated that 111 (97%) of the 115 participants in study 1 and 97 (95%) of 102 participants in study 2 recalled that product freshness was related to ingredients. It is important to note that information provided to consumers prior to product evaluation may bias them in responding in a particular, predetermined manner. In this case, the intention was to focus consumers on the intrinsic properties of the product and in turn evoke a more discriminating affective freshness response For each product type, qualitative information was analysed by word frequencies. Words cited by 10 or more participants to describe the freshness of each product type were presented (Table 4). To simplify word counts, low content ‘‘stopwords” such as ‘‘the”, ‘‘and” ‘‘with” were removed from each participants statement (ten Kleij & Musters, 2003). In many cases, different terms were associated with particular product types. For example, 72 and 66 participants considered ‘‘softness” important in the perceptions of freshness in breads and cakes respectively, but no participants used these terms to describe the freshness for biscuits (Table 4). Other terms recorded by participants as being important in one or two product types but not in the third were ‘‘crispness” (breads), ‘‘crunchy” (biscuits), ‘‘colour” (cakes), ‘‘moistness” (breads and cakes), ‘‘yeasty” (breads), ‘‘sweet” (biscuits and cakes), ‘‘sweetness” (biscuits and cakes), ‘‘malty” (biscuits and cakes), and ‘‘caramel” (biscuits and cakes). This finding suggests, that terms used to describe freshness in one product type, may in fact be considered not fresh in another product type. Important terms were represented by their common emergence between all product types in both studies such as odour, smell and buttery (Table 4). Using a similar approach to describe creaminess in different dairy products Tournier et al. (2007), showed important terms by their common occurrence across different consumer clusters. Study 2 showed similar findings to study 1, where specific terms were associated with particular product types. Apart from ‘‘crispness” in breads and ‘‘doughy” in breads and biscuits, identical terms were recorded between the two studies (Table 4). This was expected considering that these terms represented those recorded by ten or more individual participants. In this instance, words that

Table 4 Frequency of quality terms consumers most associated with fresh breads, biscuits and cakes after consumption (n = 115) and evaluating odour only (n = 102). Terms

Softness Odour Smell Flavour Aroma Crunchy Crispiness Appearance Colour Moistness Yeasty Taste Sweet Sweetness Buttery Malty Caramel Doughy

Consumption

Odour

Breads

Biscuits

Cakes

Breads

Biscuits

Cakes

72 55 24 37 19 10 21 18 13 46 34 50

– 33 13 23

66 54 19 31 21

46 97 64 26 54

– 78 51

32 78 48 17 47 12

28 49 62

11

12

12

66 15

43 15 30 46 24 34

46 11 24 15 18 23

22 67

26 15 12

30 13 16

33

41 51 44 10

35 26 31 -

occurred relatively infrequently were considered not very important and therefore ruled out any further analysis. In a similar study by Hersleth et al. (2005a), consumers used many identical or similar terms to describe perceptions for breads when terms used by more than five different consumers were represented. In addition, the strong participant agreement for the freshest products that included rye bread, brioche, croissant, butter cake, and short biscuit, in both studies, may have influenced the use of identical terms. Furthermore the testing context that questioned product ingredients in relation to freshness read out prior to product assessment may have motivated participants to focus on the perceptual attributes of freshness and therefore influenced the use of similar terms. Consumers’ have a tendency to categorise qualitative terms into either extrinsic (packaging, labelling) or intrinsic (taste, odour, texture) properties (Steenkamp & Van Trijp, 1988). Findings of this study are supported by Zeithaml (1988), who reported that following consumption consumers relied on intrinsic product properties as evaluative criteria. Hence, once these intrinsic attributes have been experienced, they can be expected to gain importance in memory. A similar investigation by ten Kleij and Musters (2003) showed that product qualities can be obtained from free-text responses that are explicitly stated by consumers. This approach has the advantage that quantitative scores as well as qualitative qualities that motivate these scores are given by the same participants. Therefore no associated meaning between product attributes and consumer freshness needs to be assumed, since descriptions of product freshness are derived by the consumers themselves. Finally, the open-ended questioning approach used in this research has the advantage of giving unbiased information, since consumers are free to express in their own words qualities that they most associated with freshness, without the influence of suggesting specific responses. 3.3. Relationships between descriptive analysis and consumer freshness Through training the assessors acquired a common qualitative and quantitative frame of reference for 28, 22 and 21 attributes that allowed the descriptive profile of breads, cakes and biscuits, respectively (Table 5). One-way ANOVA showed the descriptive attributes that significantly discriminated between different types of breads, cakes and biscuits (Table 5). Bitter taste and adhesiveness failed to significantly distinguish differences between the bread types, whilst citrus odour did not significantly discriminate between the cake types. Consequently these product attributes were not included in further analysis. All attributes significantly discriminated between the biscuit types (Table 5). In order to help identify sensory characteristics that influence consumer perceptions of freshness, PLSR analysis was used to relate descriptive sensory attributes (X-variables) with consumer freshness ratings (Y-variables) for each product type evaluated following consumption and odour evaluation only (Table 6). Results from PLSR showed calibration coefficients (strength of models) of P0.96, while validation coefficients (ability to predict new samples) were P0.83. RMSEP values ranged from 2.09 to 5.96 (on a scale from 1 to 150 mm) indicating that the current models have good predictive power. Participants that consumed breads associated freshness with a porous appearance, malty odour, sweet, buttery flavour and an oily, moist texture. Breads that were described as having a speckled appearance, grain, musty odour, coarse, hard texture and burnt after-flavour were perceived as less fresh. Similarly freshness assessed by odour only showed good relationships with positively associated attributes of sweetness and buttery flavour and negatively associated attributes of grain and musty odour (Table 6). It was apparent that freshest cakes, when consumed, were positively

Table 5 Attributes and scores (means) of baked products evaluated by descriptive sensory analysis. Attributes

Breads

Cakes

Bagel Brioche Ciabatta Croissant Foccacia Multigrain Pugliese Rye

Biscuits

PLSD Angel Butter Maderia Semolina Wholemeal F-value P-value LSD Bran Cream Malt Oat value

Short Fvalue

PLSD value

9.5

34.8

20.6

31.7

11.7

47.9

16.5

85.8 26.9

9.1

499.9 0.00

1.5

5.3

43.0

41.5

36.7

56.4

62.1

0.00

1.5

32.5

26.8

83.6

56.1 21.4

193.7 0.00

1.5

28.4

47.8

45.9

55.8

28.6

53.9

48.4

87.6 47.9

39.9

30.7

0.00

4.2

76.8

70.2

57.3

54.7

39.9

33.2

0.00

4.2

21.2

45.7

88.8

39.9 30.0

275.5 0.00

4.2

12.9

21.5

33.2

74.6

29.4

63.7

33.5

28.0 35.5

23.7

21.3

0.00

5.8

60.0 37.7

29.4 45.3

66.6 42.1

67.9 15.8 40.8 42.6

694.5 0.00 8.8 0.02

5.8 6.5

57.5 1.9

30.1 1.8

38.7 3.9

86.1 1.1

65.3 1.2

42.1 59.6

41.5 4.1

53.9 36.4 13.5 11.6

49.2 1.3

12.7 55.8

0.00 0.00

6.5 3.4

24.2

40.1

53.2

36.0 33.0

47.0

0.00

3.4

5.1 52.1

0.6 32.9

1.1 20.9

3.9 1.7 42.2 87.2

37.9 0.00 327.3 0.00

1.6 4.1

41.8 49.1 29.0

12.5 2.0 30.2

5.6 47.8 34.6

35.4 0.9 54.6 2.5 33.1 22.1

273.1 0.00 341.6 0.00 6.4 0.03

3.9 1.2 2.5

65.2 15.2 12.3

11.1 54.3 13.0

58.1 8.9 53.2

62.6 12.1 18.5 11.3 30.7 22.9

636.2 0.00 254.3 0.00 438.4 0.00

3.7 4.1 2.6

16.3 49.3

83.9 17.4

14.0 58.8

82.0 20.6

23.7 22.8

7.2 19.8

10.0 61.6

5.4 15.5 20.4 37.4

10.5 58.0

721.3 0.00 41.3 0.00

1.6 4.1

8.8 2.1 10.9 4.4

1.3 2.8 3.9 4.3

10.9 3.0 16.5 8.4

1.1 3.5 8.9 4.9

1.8 1.0 21.4 4.2

71.2 8.7 32.2 55.4

6.0 2.1 11.4 4.6

27.7 85.6 10.8 6.0

13.5 1.8 4.9 4.1

57.3 899.3 22.6 36.8

3.9 1.2 2.5 3.7

14.1 4.0 17.2 5.7

0.00 0.00 0.00 0.00

23.7 35.0 8.3 11.2

7.6 23.5 4.5 78.3

34.7 42.1 15.6 15.3

17.8 49.7 5.2 83.2

14.8 39.6 16.7 26.2

29.8 35.3 11.1 4.1

36.9 32.8 14.8 14.3

45.6 30.3 11.6 7.3

24.0 41.0 10.7 14.0

34.5 35.9 5.1 9.6

15.4 15.2 3.1 106.8

0.00 0.00 0.05 0.00

4.1 2.6 3.6 4.0

16.5 48.2 1.2 34.6 18.3 55.3 4.7

17.6 46.7 1.0 20.1 55.8 42.8 5.5

25.4 40.5 2.6 30.8 12.6 45.0 13.5

15.8 30.1 0.5 24.9 71.6 51.8 4.3

76.9 55.1 1.3 42.7 7.9 44.7 4.5

21.9 40.0 81.1 38.3 6.8 49.3 69.1

14.9 38.9 2.8 22.5 10.3 41.8 7.2

7.3 31.7 4.6 18.8 78.0 38.1 12.7

22.6 52.6 2.6 36.8 12.9 48.3 18.2

7.6 40.9 1.1 18.7 12.2 48.2 3.4

53.9 5.1 239.6 6.8 223.1 1.2 77.2

0.00 0.01 0.00 0.00 0.00 0.38 0.00

3.8 5.1 2.3 4.8 2.7 6.5 3.2

28.9 49.3 33.9 4.4 1.3 24.1

70.4 8.9 49.6 7.6 2.2 18.4

28.2 47.7 40.2 15.4 3.5 28.9

80.5 30.5 57.0 3.0 4.0 20.1

76.1 18.8 66.8 22.8 5.1 39.9

24.0 54.6 37.9 7.9 6.3 22.6

21.1 49.2 36.8 10.8 11.7 20.4

18.8 26.3 40.9 13.6 17.5 11.0

31.8 66.6 41.7 12.9 4.6 36.3

19.0 40.7 39.1 4.4 5.7 15.2

205.6 40.1 25.4 12.7 5.7 10.4

0.00 0.00 0.00 0.00 0.01 0.00

2.5 4.0 2.9 2.4 2.9 3.9

32.8 28.5

61.3 44.0

33.0 43.8

28.6 65.6

27.0 61.9

45.0 36.0

0.00 0.00

5.8 6.5

24.6 1.2 0.5 8.4

30.4 0.9 3.6 35.8

31.5 3.4 5.1 34.9

22.4 0.9 10.7 18.6

31.1 0.0 51.6 30.5

7.6 2.2 48.0 18.0

0.02 0.21 0.00 0.00

2.4 1.6 4.1 3.9

17.6 4.9

26.4 12.7

20.6 11.3

53.9 13.7

12.4 17.6

83.8 7.1

0.00 0.03

1.2 2.5

7.6 52.2

11.2 45.3

14.9 43.3

13.8 51.0

12.1 71.1

7.1 7.7

0.03 0.02

3.7 4.1

8.8 36.8

54.2 26.9

49.6 38.6

21.2 26.9

44.0 18.0

46.3 15.1

0.00 0.01

2.6 3.6

62.8

21.8

32.8

56.5 89.8

886.8 0.00

3.6

80.8 56.4

33.0 49.5

56.3 58.7

47.9 41.6

66.4 44.1

97.2 30.5

0.00 0.00

4.0 3.8

59.0 46.1

70.7 52.3

54.1 52.6

54.4 75.7 40.2 55.7

35.9 20.0

0.00 0.00

4.0 3.8

27.0 11.8 52.2 10.2 10.9

50.4 34.2 60.9 3.1 4.6

43.8 37.7 29.4 3.5 5.1

21.3 38.2 23.7 1.6 5.6

41.5 70.1 20.1 3.6 2.0

11.4 79.7 121.1 6.9 9.3

0.01 0.00 0.00 0.04 0.02

5.1 2.3 4.8 2.7 6.5

51.7 56.9 53.3 60.4

49.9 20.7 10.1 36.3

49.1 48.6 28.0 64.7

66.1 55.1 46.3 62.0

21.5 151.3 799.5 221.9

0.00 0.00 0.00 0.00

5.1 2.3 4.8 2.7

61.6

19.4

35.7

26.2

56.3

41.2

0.00

3.2

37.3

41.4

31.1

41.0 59.1

116.1 0.00

6.5

62.5 67.4 81.8 23.6

S.P. Heenan et al. / Food Quality and Preference 20 (2009) 380–390

Crumb darkness Crust darkness Crumbly Crust Smoothness Moisture Porosity Speckled Caramel Citrus Coconut Dairy Dusty Eggy Grain Malty Musty Nutty Vanilla Toasted Yeasty Bitter Buttery Doughy Oily Salty Seedy Sour Sweet Adhesiveness Coarse Crunchy Fat Hardness Moisture Bitter Burnt Sour Sweet

Sourdough White Fvalue

*

Means within a row that differ more than the corresponding least significant difference (LSD) value were significantly different (p < 0.05).

387

388

S.P. Heenan et al. / Food Quality and Preference 20 (2009) 380–390

Table 6 Results of partial least squares regression analysis showing positive and negative associations between descriptive sensory attributes and consumer freshness ratings for baked product types evaluated following consumption and smelling only odour. Product types

Consumer study

Positive attributes

Negative attributes

Calibration

Validation

RMSEP

Breads

Consumption

0.87

4.99

Odour Consumption

0.95 0.99

0.86 0.85

4.90 5.10

Cakes Biscuits Biscuits

Odour Consumption Odour

O-Caramel, O-Buttery, F-Sweet F-Sweet, T-Crunchy, T-Adhesiveness, F-Sweet O-Toasted, O-Caramel, O-Vanilla

A-Speckled, O-Musty, O-Grain, T-Coarse, T-Hardness, AFBurnt O-Grain, O-Musty, O-Dusty, F-Sour A-Crumb-darkness, A-Porosity, O-Toasted, O-Grain, OVanilla, T-Hardness AF-Burnt, AF-Bitter A-Crust-darkness, O-Malty, T-Moisture O-Musty, O-Eggy

0.96

Breads Cakes

A-Porosity, O-Malty, F-Sweet, F-Buttery, TOily, T-Moisture F-Sweet, O-Dairy, F-Buttery A-Crust-darkness, T-Moisture

0.99 0.99 0.99

0.89 0.85 0.83

4.73 4.04 2.09

*

A = Appearance, O = Odour, F = Flavour, T = Texture, AF = After-flavour.

associated with crust darkness and moist texture, whilst those evaluated by odour were associated with a caramel odour and buttery, sweet flavour (Table 6). Further to this, fresh biscuits were associated with a sweet flavour, adhesive, crunchy texture and sweet after-flavour following consumption. Biscuits evaluated by odour were positively associated with toasted, caramel, and vanilla odour. Taken collectively these results showed similarities and differences in the descriptions of freshness for the three product types tested. Fresh breads and cakes shared similar sensory qualities of sweet, buttery flavour and moist texture. In contrast fresh cakes were distinguished from fresh breads by the attribute porosity that was negatively associated to the freshness of cakes but positively associated with freshness of breads (Table 6). Moreover, fresh biscuits and cakes were similar in terms of sweet flavour and caramel odour, but were different in terms of toasted, vanilla odour and moist texture. Results from the current study showed that descriptions of freshness were dependant on the product type evaluated. This finding is supported by Cardello and Schutz (2003), who revealed that the type of food influences consumer judgements of freshness. Positively associated attributes of product freshness; moist appearance/texture, sweet, buttery flavour, malty, caramel odour, and crunchy texture (Table 6) were also reported by consumers as qualitative terms in studies 1 and/or 2 (Table 4). Possible relationships could be further drawn between appearance/colour terms used by consumers and crumb/crust darkness as described by the trained panel. In particular, consumers associated the term ‘‘buttery” with fresh breads, biscuits and cakes following either consumption or odour evaluation only (Table 4). The sensory characteristic buttery flavour was positively correlated with the freshest breads and cakes. Carr et al. (2001), showed that the greatest similarities between consumer and sensory vocabularies occurred for commonly understood attributes such as sweetness, overall flavour and crunchy/crisp texture. Comparison of the two sets of vocabularies in our study revealed that specific odour, taste and textural cues that consumers associated with product freshness could also be applied to differentiate the same products cognitively by a trained panel. This finding corroborates previous empirical evidence reported by Hersleth et al. (2005a) who demonstrated that the latent structure in consumer perceptions of seven different varieties of bread was similar to the latent structure in a trained panel’s cognitive perception of the same bread varieties. In many cases consumers will choose to fragment the meaning of complex perceptions into a combination of words, which are constructed in the form of short sentences. This makes single word counts difficult to interpret (ten Kleij & Musters, 2003). What is valuable in the current study is the discovery of which criteria are associated with freshness in each of the types of products. In particular, relating consumers’ verbal associations of bread, cake and biscuit freshness in this study with cognitive product characteristics derived from a trained panel provided a valuable insight in determining important product properties that may be the focus

for future studies. From this perspective these results bare direct relevance to product development and marketing. For example, relationships between product sensory characteristics and consumer freshness perceptions may be used to ensure newly developed baked products are perceived as fresh. In addition, product qualities that consumers’ associate with product freshness can be used by the baking industry to communicate freshness in relation to newly developed product types. It is also the case that new products are bridging what might be considered a traditional grouping of bread/cake/biscuits. For example there are now certain category of cereal bars products that have both cake and biscuit qualities. 3.4. Consumer demographics and freshness evaluations Consumer test participants who consumed baked products during the first study comprised of 55.6% female and 44.3% male. The age distribution of participants in this study was as follows: 20.8% under 25 years, 19.1% between 26 and 35 years, 18.3% between 36 and 45 years, 20.8% between 46 and 55 years, and 20.8% over 56 years. The percentage of participants in this study who consumed baked products 1–2, 3–4 and >5 times per week were 23, 44 and 33% respectively. In the second study, 57.9% female and 42.1% male participants evaluated baked product based on odour only. The age distribution of participants in this study was as follows: 20.3% under 25 years, 20.3% between 26 and 35 years, 19.2% between 36 and 45 years, 19.3% between 46 and 55 years, and 19.3% over 56 years. The results showed that 50, 25 and 25% of participants declared that they consumed baked products 1–2, 3–4 and >5 times a week respectively. The influences of age, gender and product consumption on freshness scores were analysed using ANOVA on the whole data set from each study. In study one, females’ perceived products to be significantly (p < 0.05) fresher than males (Table 7). This study supports other findings in literature that reported females giving higher ratings to product freshness (Lennernas et al., 1997; Peneau, Hoehn, Roth, Escher, & Nuessli 2006; Ragaert et al. 2004) at consumption than men. Age had a significant influence on perceived freshness (p < 0.05), with older consumers rating the products higher in terms of freshness (Table 7). The frequency of baked product consumption per week had no significant effect (p > 0.05) on product freshness. Similarly, Peneau et al. (2006) reported that the frequency of fruit and vegetable consumption did not influence the perception of product freshness. No significant difference (p > 0.05) was shown between genders in the second study when only odour was evaluated. In contrast to study 1, younger consumers aged 18–25 and 26–35 rated freshness significantly (p < 0.05) higher than older participants when only product odour was evaluated (Table 7). No significant effect (p > 0.05) was also observed for the frequency of baked product consumed per week and freshness ratings as in the first study.

S.P. Heenan et al. / Food Quality and Preference 20 (2009) 380–390 Table 7 Mean ratings of baked product freshness on a 150 mm labelled scale according to age, gender and frequency of consumption for products evaluated following consumption (n = 115) and smelling (n = 102). Demographics

Testing context consumption

Testing context odour

Age 18–5 26 35 36 45 46 55 56 65

Mean freshness rating 45.20a 44.81a 47.14ab 51.24b 50.73b F = 8.22 p < 0.001

Mean freshness rating 40.23b 41.27b 34.33a 31.35a 31.24a F = 17.99 p < 0.001

45.09b 49.47a F = 15.10 p < 0.001

36.73a 35.08a F = 1.62 p = 0.263

Frequency of consumption 1 2 46.97a 3 4 48.08a >5 48.27a F = 0.36 p = 0.699

35.22a 36.20a 36.50a F = 0.60 P = 0.551

Gender Male Female

*

Means with same letters are not significantly different (Tukey multiple comparisons test p < 0.05).

However it should be noted, that for both studies no interaction between product freshness scores, participant’s age and gender were observed. In this insistence, participants were consistent in their freshness scores across all products and it is not the case that females or older participants in study 1 and similarly younger participants in study 2 scored the freshness of specific products to be different. The present study was not designed to differentiate product freshness based on consumer demographics and consequently the observations drawn merely highlighted interesting relationships. Subsequently further studies are intended to investigate whether specific sensory drivers within particular baked product types existed among common consumer groups. 4. Conclusion This investigation demonstrated that measurements of freshness perception after either consumption or odour evaluation only in two separate consumer studies, enabled participants to very effectively discriminate between the products tested. Internal freshness mapping showed that individual consumers strongly agreed that by either consuming or smelling alone, croissant, brioche, rye bread, butter cake and short biscuit were most fresh. Qualitative information revealed that specific terms were related to particular product types as did relationships between sensory analysis and consumer freshness ratings. In particular moist appearance/texture, sweet, buttery flavour, malty odour and crunchy texture were used by both consumers and trained assessors to distinguish similarities and differences in the descriptions of freshness for the three product types. For foccacia bread, angel cake multigrain bread, madeira cake and malt biscuit, changing the mode of presentation from evaluating freshness following consumption to smelling only odour resulted in the loss of some sensory cues, which may have influence the level of perceived freshness. Consumer and descriptive sensory vocabularies from both studies were comparable, which enabled a more pertinent interpretation of the freshness perception for different baked product types that is directly relevant from the consumer perspective. This information could be applied to ensure newly developed products within each product category are perceived as fresh at the point of sale. In addition, these product qualities

389

that consumers associate with product freshness could be used by the baking industry to communicate freshness from a marketing perspective. Acknowledgements The authors wish to thank the New Zealand Foundation for Research Science & Technology (FORST) and The New Zealand Baking Industry Research Trust (BIRT) for funding this work. References Abdallah, L., Chabert, B., Le Roux, M., & Louis-Sylvestre, J. (1998). Is pleasantness of biscuits and cakes related to their actual or to their perceived sugar and fat contents? Appetite, 30, 309–324. Baardseth, P., Naes, T., & Vogt, G. (1995). Roll-in shortenings effects on Danish pastries sensory properties studied by principal component analysis. Food Science and Technology-Lebensmittel-Wissenschaft and Technologie, 28, 72–77. Brauss, M. S., Balders, B., Linforth, R. S. T., Avison, S., & Taylor, A. J. (1999). Fat content, baking time, hydration and temperature affect flavour release from biscuits in model-mouth and real systems. Flavour and Fragrance Journal, 14, 351–357. Cardello, A. V., & Schutz, H. G. (2003). The concept of food freshness: Uncovering its meaning and importance to consumers. In K. R. Cadwallader & H. Weenen (Eds.), Freshness and shelf life of foods (pp. 22–41). Washington: American Chemical Society. Carr, B. T. D., Craig-Petsinger, D., & Hadlich, S. (2001). A case study in relating sensory descriptive data to product concept fit and consumer vocabulary. Food Quality and Preference, 12, 407–412. Degel, J., & Koster, E. P. (1999). Odors: Implicit memory and performance effects. Chemical Senses, 24, 317–325. Degel, J., Piper, D., & Koster, E. P. (2001). Implicit learning and implicit memory for odors: The influence of odor identification and retention time. Chemical Senses, 26, 267–280. De Vuyst, L., & Neysen, P. (2005). The sourdough microflora: Biodiversity and metabolic interactions. Trends in Food Science and Technology, 16, 43–56. Faye, P., Bremaud, D., Teillet, E., Courcoux, P., Giboreau, A., & Nicod, H. (2006). An alternative to external preference mapping based on consumer perceptive mapping. Food Quality and Preference, 17, 604–614. Fillion, L., & Kilcast, D. (2000). Concept and measurement of freshness of fruits and vegetables. Leatherhead Food RA Research Reports No. 770. Forde, G. C., & Delahunty, C. M. (2004). Understanding the role cross-modal sensory interactions play in food acceptability in younger and older consumers. Food Quality and Preference, 15, 715–727. Frandsen, L. W., Dijksterhuis, G., Brockhoff, P., Nielsen, J. H., & Martens, M. (2003). Subtle differences in milk: Comparison of an analytical and an affective test. Food Quality and Preference, 14, 515–526. Frandsen, L. W., Dijksterhuis, G., Brockhoff, P., Nielsen, J. H., & Martens, M. (2007). Feelings as a basis for discrimination comparisons of a modified authenticity test with the same–different test for slightly different types of milk. Food Quality and Preference, 18, 97–105. Greenhoff, K., & MacFie, H. J. H. (1994). Preference mapping in practice. In H. J. H. MacFie & D. M. H. Thomson (Eds.), Measurements of food preferences (pp. 137–166). London: Blackie Academic & Professional. Guinard, J. X., Uotani, B., Mazzucchelli, R., Taguchi, A., Masuoka, S., & Fujino, S. (2000). Consumer testing of commercial lager beers in blind versus informed conditions: Relation with descriptive analysis and expert quality ratings. Journal of the Institute of Brewing, 106, 11–19. Hebeda, R. E., & Zobel, H. F. (1996). Baked goods freshness: Technology, evaluation and inhibition of staling. New York: Marcel Dekker. Heinio, R. L., Urala, N., Vainionpaa, J., Poutanen, K., & Tuorila, H. (1997). Identity and overall acceptance of two types of sour rye bread. International Journal of Food Science and Technology, 32, 169–178. Helgesen, H., Solheim, R., & Naes, T. (1997). Consumer preference mapping of dry fermented lamb sausages. Food Quality and Preference, 8, 97–109. Hersleth, M., Berggren, R., Westad, F., & Martens, M. (2005a). Perception of bread: A comparison of consumers and trained assessors. Journal of Food Science, 70, S95–S101. Hersleth, M., Ilseng, M. A., Martens, M., & Naes, T. (2005b). Perception of cheese: A comparison of quality scoring, descriptive analysis and consumer responses. Journal of Food Quality, 28, 333–349. Hummel, T., Heilmann, S., Landis, B. N., Redem, J., Frasnelli, J., Smal, D. M., et al. (2006). Perceptual differences between chemical stimuli presented through the ortho- or retronasal route. Flavour and Fragrance Journal, 21, 42–47. ISO 8586 (1994). General guidance for the selection, training and monitoring of assessors. 8586-2, I. 1994. Sensory anaylsis, International Organization for Standardization, Geneva, Switerland. Johansson, L., Haglund, A., Berglund, L., Lea, P., & Risvik, E. (1999). Preference for tomatoes, affected by sensory attributes and information about growth conditions. Food Quality and Preference, 10, 289–298. Kihberg, I., Johansson, L., Langsrud, O., & Risvik, E. (2005). Effects of information on the liking of bread. Food Quality and Preference, 16, 25–35.

390

S.P. Heenan et al. / Food Quality and Preference 20 (2009) 380–390

Koster, M. A., Prescott, J., & Koster, E. P. (2004). Incidental learning and memory for three basic tastes in food. Chemical Senses, 29, 441–453. Labbe, D., Gilbert, G., Antille, N., & Martin, N. (2009). Sensory determinants of refreshing. Food Quality and Preference, 20, 100–109. Lennernas, M., Fjellstrom, C., Becker, W., Giachetti, I., Schmitt, A., deWinter, A. M. R., et al. (1997). Influences on food choice perceived to be important by nationallyrepresentative samples of adults in the European Union. European Journal of Clinical Nutrition, 51, 8–15. Macfie, H. J. H., & Bratchell, N. (1989). Design to balance the effects of odour of presentation and first order carry-over effects in Hall tests. Journal of Sensory Studies, 4, 129–148. MacFie, H. J. H., & Hedderley, D. (1993). Current practice in relating sensory perception to instrumental measurements. Food Quality and Preference, 4, 41–49. Martens, H. (2001). Reliable and relevant modelling of real world data: A personal account of the development of PLS Regression. Chemometrics and Intelligent Laboratory Systems, 58, 85–95. Martens, H., & Martens, M. (2000). Modified Jack-knife estimation of parameter uncertainty in bilinear modelling by partial least squares regression (PLSR). Food Quality and Preference, 11, 5–16. Martens, M., Bredie, W. L. P., & Martens, H. (2000). Sensory profiling data studied by partial least squares regression. Food Quality and Preference, 11, 147–149. Martens, M., & Martens, H. (1986). Partial least squares regression. In J. R. Piggott (Ed.), Statistical procedures in food research. London: Elsevier Applied Science. Martins, Y., Pelchat, M. L., & Pliner, P. (1997). ’’Try it, it’s good and it’s good for you’’: Effects of taste and nutrition information on willingness to try novel foods. Appetite, 28, 89–102. Nordmark, B. (1999). Snack flavours revolution. Food Technology in New Zealand, 34, 26–27. Peneau, S., Brockhoff, P. B., Hoehn, E., Escher, F., & Nuessli, J. (2007). Relating consumer evaluation of apple freshness to sensory and physico-chemical measurements. Journal of Sensory Studies, 22, 313–335. Peneau, S., Hoehn, E., Roth, H.-R., Escher, F., & Nuessli, J. (2006). Importance and consumer perception of freshness of apples. Food Quality and Preference, 17, 9–19.

Pickford, M. (1999). Wheat growing, flour milling and bread baking. The NZ Trade Consortium In Association With The NZ Institute Of Economic Research (INC), Thorndon, Wellington, Working Paper No. 4. Ragaert, P., Verbeke, W., Devlieghere, F., & Debevere, J. (2004). Consumer perception and choice of minimally processed vegetables and packaged fruits. Food Quality and Preference, 15, 259–270. Rankin, K. M., & Marks, L. E. (2000). Chemosensory context effects: Role of perceived similarity and neural commonality. Chemical Senses, 25, 747–759. Richardson-Harman, N. J., Stevens, R., Walker, S., Gamble, J., Miller, M., Wong, M., et al. (2000). Mapping consumer perceptions of creaminess and liking for liquid dairy products. Food Quality and Preference, 11, 239–246. Roininen, K., Arvola, A., & Lahteenmaki, L. (2006). Exploring cosnumers’ perceptions of local food with two different qualitative techniques: Laddering and word association. Food Quality and Preference, 17, 20–30. Rozin, P. (1982). Taste-smell confusions and the duality of the olfactory sense. Perception and Psychophysics, 31, 397–401. Schutz, H. G., & Cardello, A. V. (2001). A labeled affective magnitude (LAM) scale for assessing food liking/disliking. Journal of Sensory Studies, 16, 117–159. South, G. (1989). Bread bakes a dynamic image. Supermarketing, 4, 43–45. Steenkamp, J.-B. E. M., & Van Trijp, H. C. M. (1988). Free choice profiling in cognitive food acceptance research. In D. M. H. Thomson (Ed.), Food acceptability (pp. 363–376). London: Elsevier Science Publishers. Stone, H., & Sidel, J. L. (2004). Sensory evaluation practices. Amsterdam, Boston: Elsevier Academic Press. ten Kleij, F., & Musters, P. A. D. (2003). Text analysis of open-ended survey responses: A complementary method to preference mapping. Food Quality and Preference, 14, 43–52. Tournier, C., Martin, C., Guichard, E., Issanchou, S., & Sulmont-Rosse, C. (2007). Contribution to the understanding of consumers’ creaminess concept: A sensory and a verbal approach. International Dairy Journal, 17, 555–564. Tuorila, H., Meiselman, H. L., Bell, R., Cardello, A. V., & Johnson, W. (1994). Role of sensory and cognitive information in the enhancement of certainty and liking for novel and familiar foods. Appetite, 23, 231–246. Zeithaml, V. A. (1988). Consumer perceptions of price, quality, and value – A meansend model and synthesis of evidence. Journal of Marketing, 52, 2–22.