Sensory characteristics and consumer preference for chicken meat in Guinea

Sensory characteristics and consumer preference for chicken meat in Guinea T. M. A. Sow*†1 and J. F. Grongnet*‡ *Agrocampus Ouest, Département Science et Technologie Agroalimentaires, 65 Rue de Saint-Brieuc, CS 84215, 35042 Rennes Cedex, France; †Institut Supérieur Agronomique et Vétérinaire de Faranah, Département Economie Rurale, BP 131, République de Guinée; and ‡UMR SENAH, Département AlimH., INRA, 65 Rue de Saint-Brieuc, CS 84215, 35042 Rennes Cedex, France broiler. One hundred twenty consumers expressed their preferences for the chicken samples using a 5-point Likert scale. The hierarchical cluster analysis of the preference data identified 4 homogenous consumer clusters. The hierarchical cluster analysis results showed that the live village chicken was the most preferred chicken sample, whereas the ready-to-cook broiler was the least preferred one. The partial least squares regression (PLSR) type 1 showed that 72% of the sensory data for the first 2 principal components explained 83% of the chicken preference. The PLSR1 identified that the sensory characteristics juicy, oily, sweet, hard, mouth persistent, and yellow were the most relevant sensory drivers of the Guinean chicken preference. The PLSR2 (with multiple responses) identified the relationship between the chicken samples, their sensory attributes, and the consumer clusters. Our results showed that there was not a chicken category that was exclusively preferred from the other chicken samples and therefore highlight the existence of place for development of all chicken categories in the local market.

Key words: chicken sensory characteristic, consumer preference, principal component analysis, hierarchical cluster analysis, partial least squares regression 2010 Poultry Science 89:2281–2292 doi:10.3382/ps.2010-00679

INTRODUCTION The entry of Guinea (Conakry) in the liberal system in 1984 has encouraged the Guinean economy to move toward the free market system with huge economic and financial reforms, which have resulted in entrepreneurship development, accelerating economic growth and population growth, and improvement of living conditions. Population growth has increased the consumption of animal products including chicken. To fulfill the national animal products demand growth, the Department of Agriculture, Livestock, Environment and Forestry has edited serial agricultural development ©2010 Poultry Science Association Inc. Received February 1, 2010. Accepted June 21, 2010. 1 Corresponding author: [email protected]

political letters, which have enabled the creation and development of several poultry farms in different regions of the country. In 2005, the National Direction of Livestock had estimated the poultry livestock at 16 million birds, among which 1.3 million in modern farms were composed essentially of laying hens and broilers. The poultry livestock chain is growing and is playing an important role in the food security and in the improvement of the income of the local poultry producers. This policy has allowed the country to ensure 84% of the national chicken consumption and to limit the imports at 16% of the poultry local consumption. The performance of the livestock sector in recent years has resulted in an annual growth of the poultry livestock of around 10% and a contribution of up to 4.5% to the gross domestic product. The success of this agricultural policy, since the 1980s, has shown significant results

2281

Downloaded from https://academic.oup.com/ps/article-abstract/89/10/2281/1500984 by guest on 23 January 2019

ABSTRACT This study identified the sensory characteristics and consumer preference for chicken meat in Guinea. Five chicken samples [live village chicken, live broiler, live spent laying hen, ready-to-cook broiler, and ready-to-cook broiler (imported)] bought from different locations were assessed by 10 trained panelists using 19 sensory attributes. The ANOVA results showed that 3 chicken appearance attributes (brown, yellow, and white), 5 chicken odor attributes (oily, intense, medicine smell, roasted, and mouth persistent), 3 chicken flavor attributes (sweet, bitter, and astringent), and 8 chicken texture attributes (firm, tender, juicy, chew, smooth, springy, hard, and fibrous) were significantly discriminating between the chicken samples (P < 0.05). Principal component analysis of the sensory data showed that the first 2 principal components explained 84% of the sensory data variance. The principal component analysis results showed that the live village chicken, the live spent laying hen, and the ready-to-cook broiler (imported) were very well represented and clearly distinguished from the live broiler and the ready-to-cook

2282

Sow and Grongnet

MATERIALS AND METHODS Chicken Description The chicken supply in Guinea consists of 3 types of chickens as they are produced by i) households living

in villages (live village chickens), ii) modern local farms (live broilers, ready-to-cook broilers, and live spent laying hens), or iii) they are imported from abroad (readyto-cook broilers). Generally speaking, live village chickens are indigenous chickens. Their genetics are of local origin with, sometimes, an old infusion of European strain in an imprecise and low proportion. They are raised in a very traditional way, without regular feeding from the villagers. They ramble around the houses to get their feed. They are not sold at a precise age but when their owners need some extra money because, for the villagers, such birds represent a type of savings to face special situations (e.g., sickness, scholar fees, and tuition). Live spent laying hens, live broilers, and ready-to-cook broilers are raised in modern farms. They are genetically modern and diverse, some used strains being specially adapted to tropical conditions. With the exception of the climate, feeding, age at sale-slaughtering for broilers, or end of laying cycle for hens are similar to American or European farming conditions. Among these 3 types of chicken, only ready-to-cook broilers (imported) and a small part of live broiler production are sold after being slaughtered. The live village chicken is sold alive in villages or in various markets in the country. Compared with the categorization of the market supply, we conducted our study by subdividing the poultry market taking into account 3 criteria: the state in which the chicken is purchased (slaughtered or not), the place of purchase (village market, farm market, or supermarket), and the farming method (traditional or modern). Thus, 5 types of chicken were purchased and used to conduct this study. Samples were divided into 5 categories according to the criteria mentioned above and are described in Table 1. These types are the main chicken categories sold on the market and therefore, they are representative of the chicken local market supply. Chicken samples for the consumer preference study were similar to those of the sensory analysis.

Chicken Sample Preparation The preparation of the chicken samples used in this study was conducted as follows: live village chickens and farm chickens (live broilers and live spent laying hens) were purchased directly from village and farm markets, respectively, before being slaughtered according to the traditional rite with a knife. Slaughtered chickens were then plucked, gutted, vacuum-packed, and stored in a refrigerator at 4°C (Lawlor et al., 2003; Nantachai et al., 2007). Ready-to-cook broilers and ready-to-cook broilers (imported) were purchased in supermarkets before being transported and also stored at 4°C (Nantachai et al., 2007). Before the roasting, the chickens were coated with shea butter and seasoned with a mixture of salt, onion, pepper, bay leaves, and sweet pepper. The roasting was done with an electric roaster rented from a professional manufacturer of roast meat. Roast-

Downloaded from https://academic.oup.com/ps/article-abstract/89/10/2281/1500984 by guest on 23 January 2019

and unveiled, at the same time, some weaknesses that are challenges that require now a new orientation of the agricultural production in general and that of livestock in particular to provide more processed products in a way to meet consumer hopes and needs. The country’s increasing openness to the outside, the globalization of food trade, and the involvement of women in business promote changes in lifestyle and eating habits of populations. These changes result in increased demand and consumption of processed chicken (i.e., slaughtered and ready-to-cook products). Unfortunately, the Guinean poultry industry’s weaknesses are 3-fold. First, a large part of the national poultry consumption is provided by the traditional poultry producers without added value for consumers and requires more time for preparation because these chickens are sold alive. Second, the modern poultry sector’s processed chicken supply is weak and therefore fails to respond efficiently to the Guinean consumer’s demand for processed poultry products. Third, the increasing implication of Guinean women in economic life and business is dramatically limiting the food cooking time, leading to switching behavior in favor of imported poultry products and establishing, at the same time, a challenge and a threat to the local poultry sector. Challenges and threats from trade globalization have raised needs for studying the sensory profile and preference of Guinean chicken consumers for the purpose of government policymakers and local poultry producers. Many authors have showed that sensory analysis allows producers to identify, understand, and respond to consumer preferences more efficiently (Hashim et al., 1995; McEwan, 1996; Owens and Sams, 1998; Liu et al., 2004; Fanatico et al., 2007; Saha et al., 2009). Furthermore, the identification of sensory characteristics and consumer preference helps industry producers to segment their market and to increase their competition strengths (Tan et al., 2001; Lawlor et al., 2003; Ponte et al., 2004; Young et al., 2004). Thus, for the first time to our knowledge, within a special Guinean context, this study attempts to analyze the sensory characteristics and consumer preferences for chicken meat to respond to the following 3-fold objective: i) identifying the sensory characteristics and consumer preference for chicken categories sold on the local market, ii) segmenting consumers to determine segments with similar preferences, and (iii) highlighting the sensory characteristics that determine consumer segments’ preference for each category of chicken. The results of this study will undoubtedly help to respond to the needs of government policymakers and local poultry supply chain producers.

SENSORY CHARACTERISTICS AND CONSUMER PREFERENCE FOR CHICKEN MEAT

2283

Table 1. Chicken sample characteristics, codes, and descriptions State of purchase

Place of purchase

Live village chicken

Traditional

Not slaughtered

Village

Live broiler

Modern

Not slaughtered

Live spent laying hen

Modern

Ready-to-cook broiler Ready-to-cook broiler (imported)

Carcass average weight (g)

Code

Description

1,202.2

A

Farm

1,502.4

B

Not slaughtered

Farm

1,006.6

C

Modern

Slaughtered

Supermarket

1,601.8

D

Modern

Slaughtered

Supermarket

1,832.4

E

Chicken raised traditionally by households living in villages Chicken raised in modern local farms and sold alive Chicken raised in local farms for egg production and sold after laying period Chicken raised in local farms for chicken meat and sold processed Chicken imported from other countries

ing was done in the Aid for Youth Non Governmental Organization facility (Sangoyah, Conakry) by a professional in accordance with our recommendations and study requirements. The thermometer and timer of the roaster were used to set the cooking temperature to 160°C and the cooking time to 1 h 15 min. Of course, cooking quality control was monitored throughout roasting to ensure the proper cooking of chicken inside and to offer an appropriate taste to assessors. The roasted chickens were then cooled before being cut into small cubic pieces, without skin, and placed in transparent glass plates with covers. The samples, composed of a mix of light and dark meat cubic pieces, were then coded with letters (Table 1) before being submitted to sensory testers. The Guinean preference for chicken meat color is very heterogeneous and therefore varies very significantly between consumers. The chicken samples for the consumer preference test were treated similarly before being submitted to consumers for the hedonic scoring (Lawlor et al., 2003; Thompson et al., 2004; Nantachai et al., 2007).

Test Room Conditions and Environment Given the novelty of such a study in Guinea and climate conditions in tropical areas, we completed this study in the refectory of the Aid for Youth NGO facility, which qualifies for room temperature through its air conditioning equipment. The day before each test, the refectory was washed, disinfected, and deodorized and dining tables and chairs were cleaned to create a pleasant atmosphere and thus limit the influence of sensations of the external environment on the natural ability of an assessor’s judgment. Furthermore, we took care to carry out this experience during weekends to minimize the influence of noises and odors and other forms of disturbances during the test sessions (Lawlor et al., 2003; Thompson et al., 2004). The inside of the test room was divided into 2 parts. The first part served as a meeting place for preparation of the test and the second part served as a private place for evaluating products. The private place was divided into 10 booths with opaque cardboard to allow each assessor to isolate himself or

herself during the evaluation. Each booth was equipped with a table, a chair, and a pencil (Lawlor et al., 2003; Thompson et al., 2004).

Assessor Selection and Training We conducted our experiment in accordance with the international standards for selection, training, conducting, and designing a room test for sensory analysis (ISO 8586, ISO, 1993; ISO 6658, ISO, 2005; ISO 8589, ISO, 2007). Human senses are truly a chemistry laboratory for identifying every type of sensation, but they are not, every time, able to distinguish or make a clear difference between sensations (Boughter and Bachmanov, 2008; Wise et al., 2008). Because of that, any study of sensory analysis must begin by some training sessions that can help assessors discuss, understand, and select a correct vocabulary to conveniently describe the product to study. We started by recruiting 10 people based on their chicken consumption frequency (at least once a week) and sociodemographic characteristics (Lawlor et al., 2003). Panelists were 5 men and 5 women. Due to ethnical food habit differences, as a precaution, we decided to balance the representation of ethnic groups and to involve in the study 2 Fular, 2 Soso, 2 Maninka, 2 Kissi, and 2 Toma. These ethnic groups are the major components of the Guinean population. Moreover, the average age for men was 20.8, whereas women were younger than men, with an average age of 19.4. All panelists were chemistry students at the University of Conakry. Relying on these criteria, those identified people were trained for 3 d on chicken sensory study. Each day, the training session took 2 h, during which candidates discussed the characteristics of the 5 categories of chicken to be studied. These discussions fulfilled 3 main objectives, namely: i) knowledge deepening and differentiation of all chicken categories according to their characteristics (Table 1); ii) identification, definition, and familiarization with the sensory attributes; and iii) the selection of a final list of terms to be used in the chicken sensory study. These training sessions led to a greater understanding of the chicken sensory attributes, which were not previ-

Downloaded from https://academic.oup.com/ps/article-abstract/89/10/2281/1500984 by guest on 23 January 2019

Name

Farming method

2284

Sow and Grongnet

Descriptive Sensory Analysis Three consecutive days were chosen for conducting the sensory test. During each session, each candidate was provided with toothpaste and a toothbrush to prepare a short time before the beginning of the test. The assessors were then gathered around a table. A small reminder was made about all attribute definitions and the importance of the study and the needed

seriousness for providing scores reflecting natural feelings was discussed. Assessment forms were distributed and the assessors were reminded of the scale. Chicken sample presentation was monadic and simultaneous for all samples and for all assessors. After receiving the samples, each candidate isolated himself or herself to evaluate the product for 20 min. Based on the Likert scale discussed during the training sessions, candidates assessed each sample from 1 to 5 according to their level of appreciation of the expression of each descriptor. The evaluation forms were left in the booth and then recovered to be stored in a binder prepared for this purpose. After each sample test, candidates were asked to cleanse their palate with mineral water called Coyah Yé before the following sample test to remove residual feeling and prevent adaptation (Delwiche and O’Mahony, 1996). Coyah is a town historically habited by the Soso ethnic group. The factory producing the mineral water is implanted in this town and Yé means water in Soso, the habitants’ language. The following tests were organized similarly. The last 2 d of tests were organized similarly and each took 2 h 45 min.

Consumer Preference Assessment Overall preference assessment for chicken categories was carried out by a sample of 120 people. Participants were men (47%) and women (53%) more than 20 yr old with a high level of education (Aid for Youth NGO workers and students). Ethnically, the consumer sample was representative of the Guinean population with approximately 20% for each major ethnic group. The selection of the participants was, in addition, based on their chicken consumption habits (Lawlor et al., 2003; Carbonell et al., 2008). Only people who consumed chicken at least twice per month were selected for chick-

Table 2. Sensory characteristics used in the study with their codes and definitions Sensory characteristic

Code

Definition

Appearance   Brown   White   Yellow Odor   Oily   Intense   Medicine smell   Roasted   Mouth persistent Flavor   Sweet   Bitter   Astringent Texture   Firm   Tender   Juicy   Chewy   Smooth   Springy   Hard   Fibrous

  Br Wh Ye   Oi In Ms Ro Mf   Sw Bi As   Fi Te Ju Ch Sm Sp Ha Fb

  The intensity of the brown color of the meat The intensity of the white color of the meat The intensity of the yellow color of the meat   The extent of the fat odor of the meat The extent of the chicken odor Smell associated with chemicals The extent of the smell of the roast Taste remaining in the mouth after chewing and swallowing the meat   Intensity of sweetness taste on the tongue Intensity of the bitter taste on the tongue Intensity of the meat to dry the mouth

 

The degree of chicken meat resistance to teeth pressure associated with a dryness feeling The intensity of the tenderness of the chicken meat The ability of chicken meat to produce juice in the mouth Ease of chewing the chicken meat after the first bite The level of the softness of the chicken meat between teeth The degree of elasticity of the chicken meat The extent to which the chicken meat requires efforts when chewing The fibrous nature of the chicken meat found when masticating

Downloaded from https://academic.oup.com/ps/article-abstract/89/10/2281/1500984 by guest on 23 January 2019

ously clearly distinguished by the candidates. Through these training sessions, the assessors identified the sensory attributes for describing chicken appearance, odor, flavor, and texture. On the last training day, 30 min of the 2-h session were devoted to the revision and development of the final list of sensory attributes. Thus, from the training sessions, there were 4 identified sensory descriptors for appearance, 6 descriptors for odor, 6 descriptors for flavor, and 9 sensory descriptors for texture. For names, codes, and definitions of sensory attributes, see Table 2. At the end of the last training session, panelists selected 25 sensory attributes to be used for the study. However, it is important for comprehension to note that only the attributes that significantly discriminated between chicken samples were presented in the study (Table 2). The sensory attributes that did not discriminate between chicken samples were withdrawn from the subsequent analysis (bright, blood-smelling, fermentedsmelling, acidic, salty, and sticky). The Likert 5-point psychometric scale was discussed during the training sessions and comprehensive vocabulary was selected for the chicken sensory assessment (Lawlor et al., 2003; Carbonell et al., 2008). The scale ranged from 1 (low expression of the attribute) to 5 (high expression of the attribute).

SENSORY CHARACTERISTICS AND CONSUMER PREFERENCE FOR CHICKEN MEAT

en preference evaluation (Gou et al., 1998; Lawlor et al., 2003). Given the difficulty for gathering a large number of people at the same time and at the same place and a cost reduction objective, we made a single test. The test was conducted in the refectory of the Aid for Youth NGO facility. Test room conditions and environments were similar to those for the sensory analysis (Lawlor et al., 2003). Chicken preference rating was based on the Likert 5-point scale ranging from 1 (I do not like at all) to 5 (I like very much). Sample presentation was similar to that of the sensory analysis experiment.

The sensory and consumer preference data were saved in an Excel spreadsheet (Microsoft Corp., Redmond, WA). To identify the sensory attributes that discriminate between chicken samples, rating scores mean for each sensory attribute across assessors was calculated and then analyzed by means of ANOVA (ANOVA type I). Sensory attributes that did not distinguish between the chicken samples were withdrawn for further study (P > 0.05). A Tukey’s honestly significant difference (HSD) post hoc test was then carried out to determine sensory attributes means, which significantly differ for the chicken samples (Lawlor et al., 2003; De Marchi et al., 2005; Nantachai et al., 2007). Significant sensory attributes were then standardized (1/SE) to allow comparison between them (Lawlor et al., 2003; GallardoEscamilla et al., 2005). A principal component analysis (PCA) was then conducted on the standardized data (Lawlor et al., 2003; Gallardo-Escamilla et al., 2005; Carbonell et al., 2008). Analysis of variance type I was then performed on the generated principal components (PC) for identifying those that were significant (Liu et al., 2004; Gallardo-Escamilla et al., 2005). The first 2 significant PC were then chosen for result plotting and interpretation (Liu et al., 2004; Chapman et al., 2005; Gallardo-Escamilla et al., 2005). The consumer preference data were also standardized (1/SE) before being analyzed by means of PCA (Lawlor et al., 2003; Liu et al., 2004; Gallardo-Escamilla et al., 2005). Consumers were assimilated as variables and chicken samples were considered as individuals (Gallardo-Escamilla et al., 2005). To segment consumers, hierarchical cluster analysis (HCA) with squared Euclidian distances and Ward’s method were carried out on the consumer chicken preference scores (Carbonell et al., 2008). Finally, 2 partial least squares regression models (PLSR) were built for studying the relationship between sensory characteristics and consumer preference (Lawlor et al., 2003; Chapman et al., 2005; GallardoEscamilla et al., 2005). The PLSR1 was carried out by regressing the average preference score for all consumers onto the sensory attributes for identifying the relevant chicken sensory attributes that drive the chicken preference of Guinean consumers. The PLSR2 was done by regressing the sensory characteristics onto consumer

preference clusters (Lawlor et al., 2003; Gallardo-Escamilla et al., 2005). The R software release 2.8.0. for statistical computing and graphics was used to carry out all of the basic statistical analysis (R Development Core Team, 2008), whereas sophisticated analysis and modeling were done by using The Unscrambler software version 9.8 (CAMO Software AS, Oslo, Norway).

RESULTS AND DISCUSSION Descriptive Sensory Analysis As we reported earlier, after having recorded the sensory data in an Excel spreadsheet, we computed the mean of each sensory characteristic across all assessors and performed an ANOVA type I to identify the sensory attributes that discriminate between the chicken samples (Table 3). The ANOVA results showed that 3 chicken appearance attributes (brown, yellow, and white), 5 chicken odor attributes (oily, intense, medicine smell, roasted, and mouth persistent), 3 chicken flavor attributes (sweet, bitter, and astringent), and 8 chicken texture attributes (firm, tender, juicy, chew, smooth, springy, hard, and fibrous) were significantly discriminating between the chicken samples and were kept for the rest of the study. In addition, the results of the ANOVA showed that panelists were more able to distinguish between the chicken samples by their appearance and texture than by their odor and flavor. Lawlor et al. (2003) found the same results when studying the sensory attributes and consumer preference for chicken in Ireland. In contrast to our results and those found by Lawlor et al. (2003), Nantachai et al. (2007) found that Thai chicken assessors were not able to discriminate between chicken samples using their flavor attributes. An ANOVA also showed that there was no assessor and session effects (P > 0.05) on the chicken samples. Further analysis was also investigated to find sensory attributes’ means that discriminate between the chicken samples by performing the Tukey’s HSD pairwise comparison test (Young et al., 2004; Nantachai et al., 2007). The probability of detecting the HSD was set at the 5% probability level (P < 0.05). Tukey’s test is suitable for comparing means of samples with equal size (Nantachai et al., 2007). The significant sensory data were then standardized and analyzed by means of PCA (Murray and Delahunty, 2000; Lawlor et al., 2003; Gallardo-Escamilla et al., 2005). The PCA results showed that the first 2 PC significantly discriminated between the chicken samples (P < 0.05) and explained 84% of the sensory data variation. The PCA results showed that the chicken types A, C, and E were very well represented and clearly distinguished from the others 2 chicken categories, B and D (Figure 1). The first PC, which explained 48% of the variance in the sensory data, was predominantly described by the ready-to-cook broiler (imported) (E) and the traditional live village chicken (A). As it can

Downloaded from https://academic.oup.com/ps/article-abstract/89/10/2281/1500984 by guest on 23 January 2019

Data Analysis

2285

2286

Sow and Grongnet

Table 3. Mean scores of the chicken samples’ sensory characteristics with corresponding results of 1-way ANOVA and Tukey’s honestly significant difference (HSD) test Sample code1 Sensory characteristic  

 

 

 

2.46b 2.66b 4.26a 3.33a 3.23ab 1.86b 2.76ab 3.50a 3.86a 2.73ab 2.03c 3.06b 3.03ab 3.56a 2.26b 2.23b 2.13bc 3.46a 2.36c

B  

 

 

 

3.23ab 2.50b 2.10c 2.80ab 3.90ab 2.33b 2.36b 3.26a 3.13ab 2.36b 3.40ab 2.60b 2.46bc 2.66b 3.23a 3.13a 2.76ab 3.26a 3.23ab

C  

 

 

 

3.86a 2.43b 2.26bc 2.63bc 3.13b 2.16b 3.46a 3.10a 3.06b 3.46a 4.00a 4.10a 1.80c 1.93c 3.33a 3.16a 1.90c 3.10a 3.86a

D  

 

 

 

3.06ab 2.83b 2.23bc 2.00cd 3.56ab 3.20a 2.56b 2.76a 3.00b 2.56b 2.93b 2.73b 2.63b 2.80b 3.23a 2.93ab 2.06bc 2.76a 3.06bc

E  

 

 

 

3.10ab 4.26a 2.90b 1.63d 3.93a 3.93a 2.36b 1.36b 2.63b 2.36b 3.20ab 3.06b 3.53a 2.53bc 3.73a 3.63a 3.30a 1.36b 3.03bc

F-value  

 

 

 

5.24 20.21 21.72 16.40 3.58 22.57 5.72 21.79 6.39 5.61 11.46 10.96 13.03 15.26 6.93 6.48 9.57 21.14 9.01

P-value  

 

 

 

0.001 0.000 0.000 0.000 0.008 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000

HSD  

 

 

 

0.89 0.69 0.79 0.67 0.79 0.73 0.78 0.74 0.73 0.78 0.87 0.72 0.73 0.61 0.84 0.82 0.77 0.74 0.73

a–dMeans

within a row with the same superscript are not significantly different depending on the results of Tukey’s test. = chicken raised traditionally by households living in villages; B = chicken raised in modern local farms and sold alive; C = chicken raised in local farms for egg production and sold after laying period; D = chicken raised in local farms for chicken meat and sold processed; E = chicken imported from other countries. 1A

be seen in Figure 1, the PCA results show a clear opposition between the live village chicken (A) and the ready-to-cook broiler (imported) (E). The second PC, accounting for 36% of the variation in the sensory data, was mainly described by the live spent laying hen (C). The ready-to-cook broiler (imported) (E) was highly correlated to the first PC (r = 0.99) and was characterized by the high expression of the sensory attributes white, chewy, smooth, medicine smell, intense odor, springy meat, and tender meat. These sensory characteristics of the ready-to-cook broiler (imported) reflect the young age at which this broiler is slaughtered (Northcutt et al., 2003). The live village chicken (A) was highly and negatively correlated to PC1 (r = −0.97) and was characterized by the high expression of the sensory attributes yellow, hard, oily, mouth persistent, sweet, and juicy. The description of the live village chicken (A) by these sensory attributes highlights the old age at which this bird is slaughtered and its farming method. Nantachai et al. (2007) and Jaturasitha et al. (2008) found that Thai indigenous chicken is characterized by the same sensory attributes as those found for the live village chicken (A) in Guinea. The live spent laying hen (C) was negatively correlated to PC2 (r = −0.99). Unlike the live village chicken (A), the live spent laying hen (C) was assessed by panelists as brown meat color, bitter, firm, fibrous, astringent, and roast odor. The association of the live spent laying hen with these sensory attributes shows that this bird is mainly described as a bird with a poor sensory quality. This result is not surprising because the live

spent laying hens are old birds with depleted fat tissue (Leclerq and Simon, 1982). Unlike the 3 previous samples, the chicken samples B and D were very close but not well represented and described by the panelists. This result is not surprising because the only difference between these 2 chickens is their state of purchase. The live broiler (B) was purchased alive, whereas the readyto-cook broiler (D) was purchased slaughtered from the supermarket. Because these 2 chicken samples are very close to each other and not well represented, it appears that they have not been predominantly characterized by any sensory attribute. In brief, it appeared that panelists were more able to differentiate the live village chicken (A), live spent laying hen (C), and ready-to-cook broiler (imported) (E) than the ready-to-cook broiler (D) and live broiler (B).

Consumer Preference Assessment As pointed out previously, the consumer chicken preference assessment was carried out under the same conditions as those for the sensory evaluation. The means and SD of the 5 chicken samples preference data are presented in Table 4. From this result, the first thing to note is the very similar scores obtained by the 5 types of chicken, the smallest mean being 3.44 for the ready-to-cook broiler (D) and the biggest being 3.87 for the live village chicken (A) (Figure 2). The preference scores were standardized and analyzed by means of PCA (Murray and Delahunty, 2000; Lawlor et al., 2003;

Downloaded from https://academic.oup.com/ps/article-abstract/89/10/2281/1500984 by guest on 23 January 2019

Appearance   Brown   White   Yellow Odor   Oily   Intense   Medicine-smelling   Roasted   Mouth persistent Flavor   Sweet   Bitter   Astringent Texture   Firm   Tender   Juicy   Chewy   Smooth   Springy   Hard   Fibrous

A

2287

SENSORY CHARACTERISTICS AND CONSUMER PREFERENCE FOR CHICKEN MEAT

Table 4. Results of the hierarchical cluster analysis of the consumer preference data showing homogenous consumer clusters, 1-way ANOVA, and least significant difference (LSD) test results Chicken sample1,2 Cluster 1 2 3 4 F-value P-value LSD All

Cluster size 18 43 20 39 — — — 120

A 3.55b,3 4.30a,4 1.80c,3 4.58a,4

45.58 0.000 0.50 3.874 (1.34)

B 3.94a 3.16b 3.95a,4 4.38a

8.92 0.000 0.59 3.81 (1.19)

C 4.83a,4 2.67c,3 3.70b 3.61b

18.13 0.000 0.59 3.48 (1.28)

D 4.77a 4.13b 3.25c 2.15d,3

49.21 0.000 0.49 3.443 (1.33)

E 4.38a 3.74ab 2.90c 3.07bc

6.61 0.001 0.68 3.48 (1.33)

within a column with the same superscript are not significantly different according to the LSD test results. = chicken raised traditionally by households living in villages; B = chicken raised in modern local farms and sold alive; C = chicken raised in local farms for egg production and sold after laying period; D = chicken raised in local farms for chicken meat and sold processed; E = chicken imported from other countries. 2The SD of the overall means of each chicken sample is in parentheses. 3The least preferred chicken sample. 4The most preferred chicken sample. 1A

Gallardo-Escamilla et al., 2005) using The Unscrambler software version 9.8 (CAMO Software AS). In the PCA, consumers were considered as the variables and chicken samples were assumed as the indi-

viduals (Gallardo-Escamilla et al., 2005). The results of the PCA showed that the first 2 PC explained 62% of the consumer preference data variance. Through the examination of the consumer preference map, it ap-

Figure 1. Principal component analysis results showing the description of the 5 chicken categories by the 19 sensory attributes. A = chicken raised traditionally by households living in villages; B = chicken raised in modern local farms and sold alive; C = chicken raised in local farms for egg production and sold after laying period; D = chicken raised in local farms for chicken meat and sold processed; E = chicken imported from other countries. Br = the intensity of the brown color of the meat; Wh = the intensity of the white color of the meat; Ye = the intensity of the yellow color of the meat; Oi = the extent of the fat odor of the meat; In = the extent of the chicken odor; Ms = smell associated with chemicals; Ro = the extent of the smell of the roast; Mf = taste remaining in the mouth after chewing and swallowing the meat; Sw = intensity of sweetness taste on the tongue; Bi = intensity of the bitter taste on the tongue; As = intensity of the meat to dry the mouth; Fi = the degree of chicken meat resistance to teeth pressure associated with a dryness feeling; Te = the intensity of the tenderness of the chicken meat; Ju = the ability of chicken meat to produce juice in the mouth; Ch = ease of chewing the chicken meat after the first bite; Sm = the level of the softness of the chicken meat between teeth; Sp = the degree of elasticity of the chicken meat; Ha = the extent to which the chicken meat requires efforts when chewing; Fb = the fibrous nature of the chicken meat found when masticating. PC1 = principal component 1; PC2 = principal component 2.

Downloaded from https://academic.oup.com/ps/article-abstract/89/10/2281/1500984 by guest on 23 January 2019

a–dMeans

2288

Sow and Grongnet

pears that there was a high concentration of consumers around the live village chicken (A) and the live broiler (B), meaning that a large group of consumers in the sample expressed their preference for these 2 types of chicken (Murray and Delahunty, 2000; Young et al. 2004). Also, it was clear that the ready-to-cook broiler (D) and ready-to-cook broiler (imported) (E) received a high liking score from consumers compared with the preference score of the live spent laying hen (C). The first PC, which explained 33% of the preference data variance, showed preference opposition between the traditional live village chicken (A) (positively correlated to PC1, r = 0.99), the live broiler (B) (negatively correlated to PC1, r = −0.75), and the live spent laying hen (C) (negatively correlated to PC1, r = −0.99). The second PC, which explained an additional 29% of the preference data variance, opposed the live chickens (live village chicken and live broiler) to the processed chickens [ready-to-cook broiler (negative correlation to PC2, r = −0.85) and ready-to-cook broiler (imported) (negative correlation to PC2, r = −0.48)]. The chicken sensory attributes projection on the sensory map (Figure 1) showed that consumers liked the live village chicken (A) because of its yellow color, juicy meat, sweet meat, oily meat, hard meat, and mouthpersistent odor, whereas the live spent chicken (C) was appreciated because of its brown color, astringent meat, fibrous meat, and firm meat. The ready-to-cook broiler (imported) (E) was liked by consumers whose sensory

profile was characterized by sensory attributes such as springy, white meat, intense odor, smooth meat, and medicine smell. Briefly, the internal preference mapping showed that consumers involved in this study expressed a high preference score for the live village chicken (A) and the live broiler (B). The concentration of consumers on the internal preference map showed also that the live spent laying hen (C) was less preferred than the ready-to-cook broiler (D) and the ready-to-cook broiler (imported) (E). It is recognized that average values obtained from consumer preference ratings can help to identify general tendencies about the preference for products but cannot help to determine a group of people that can prefer some products over others and vice versa (Carbonell et al., 2007). Because of this lack of information of the internal preference mapping, we performed a further analysis using a segmentation technique for identifying similarities among consumers to segment the chicken market. Different techniques have been previously used in consumer segmentation (Carbonell et al., 2007), including the HCA (McEwan,1998; Vigneau et al., 2001; Santa Cruz et al., 2002; Jaeger et al., 2003).

HCA of the Consumer Preference Data Hierarchical cluster analysis is a statistical method for finding relatively homogeneous clusters of consumers based on measured preference (McEwan, 1998;

Downloaded from https://academic.oup.com/ps/article-abstract/89/10/2281/1500984 by guest on 23 January 2019

Figure 2. Principal component analysis results showing the preference scores of 120 consumers for the 5 chicken categories. A = chicken raised traditionally by households living in villages; B = chicken raised in modern local farms and sold alive; C = chicken raised in local farms for egg production and sold after laying period; D = chicken raised in local farms for chicken meat and sold processed; E = chicken imported from other countries. PC1 = principal component 1; PC2 = principal component 2.

SENSORY CHARACTERISTICS AND CONSUMER PREFERENCE FOR CHICKEN MEAT

Relationship Between Consumer Preference and Chicken Descriptive Sensory Data The relationship between chicken consumers’ overall preference score and chicken sensory characteristics was investigated using a PLSR1 (single response; Murray and Delahunty, 2000; Gallardo-Escamilla et al., 2005). This investigation was carried out to identify the most relevant chicken sensory attributes that underpinned the Guinean chicken consumer preference. The results of the regression model are presented in Figure 3. How the model fitted the data was investigated by examining the correlation coefficient (r). The model correlation coefficient (r = 0.68) showed that the PLSR1 predicted

fairly well the chicken preference. The PLSR1 results showed that 72% of the sensory data for the first 2 PC explained 83% of the consumer preference data. The PLSR1 results indicated that the chicken sensory characteristics juicy, oily, sweet, hard, mouth persistent, and yellow color were highly correlated with the consumer preference and therefore were considered as the most relevant sensory drivers of the Guinean chicken preference (Figure 3). Unlike the attributes described above, the other sensory attributes were not considered as relevant sensory characteristics of the Guinean chicken consumption. Lawlor et al. (2003) and Nantachai et al. (2007) also found that traditional chickens were not more preferred than modern chickens in studying the sensory characteristics and consumer preference for chicken meat in Ireland and Thailand, respectively.

External Preference Mapping of the Descriptive Sensory Attributes and Consumers’ Cluster Preference A PLSR2 (multiple responses) model was used for investigating the relationship between chicken samples, their sensory characteristics, and consumer cluster preference (Hough and Sanchez, 1998; Murray and Delahunty, 2000; Lawlor et al., 2003). Partial least squares regression is a multivariate method allowing one to relate the variation in one (PLSR1) or several response (PLSR2) variables to the variation of several predictors, with explanatory or predictive purposes (Gallardo-Escamilla et al., 2005). In this section, we used the PLSR2 model for describing the chicken preference of the consumer clusters obtained from the hierarchical clustering of the preference data. In the model, the clusters of the consumers were considered as the variables to be predicted and the sensory attributes as the explanatory variables (Murray and Delahunty, 2000; Lawlor et al., 2003). The obtained results of the PLSR2 were an external preference mapping (Gou et al., 1998; Murray and Delahunty, 2000; Lawlor et al., 2003) showing the graphical visualization of the association between the chicken samples, their sensory attributes, and the consumer clusters. The results of the model showed that 85% of the sensory data for the first 2 PC explained 67% of the clusters’ chicken preference (Figure 4). The ability of the model to fit the data was assessed by its correlation coefficient, which was relatively high (r = 0.70). In addition, the ability of the model to predict the preference of each cluster was examined through the correlation coefficient (r; Murray and Delahunty, 2000). The analysis of the projection of each consumer cluster showed that cluster 1 (r = 0.69) and cluster 3 (r = 0.70) were very associated with the live spent laying hen (C) and the ready-to-cook broiler (imported) (E). The chicken samples B and D were less associated with the consumer clusters 1 and 3. These groups of consumers were influenced by sensory attributes such

Downloaded from https://academic.oup.com/ps/article-abstract/89/10/2281/1500984 by guest on 23 January 2019

Thompson et al., 2004; Lee et al., 2009). The HCA carried out on the consumer preference data allowed us to identify 4 clusters of consumers with homogenous chicken preference (Table 4). The HCA results showed that the most preferred chicken sample was the live village chicken (A). Likewise, the HCA analysis revealed that consumers’ least liked chicken sample was the ready-to-cook broiler (D). However, the examination of the preference score within each of the 4 clusters showed some preference differences between all chicken samples. Thus, cluster 1, the smallest segment of consumers with 15% of the consumer sample, showed the highest preference score for the live spent chicken (C), whereas its least liked chicken sample was the traditional live village chicken (A). Cluster 2, the largest cluster, which contained 36% of the chicken sample, showed the highest preference score for the live village chicken (A). Cluster 2 particularly disliked the live spent laying hen (C). Cluster 3 consisted of 17% of the consumer sample, showed high preference for the live broiler (B), and disliked the live village chicken (A). Cluster 4, the second largest cluster, contained 33% of the consumer sample and also preferred the live village chicken (A), whereas it particularly disliked the ready-to-cook broiler (D). The HCA showed that the ready-to-cook broiler (imported) (E) was averagely appreciated. The consumer population preference distribution through the clusters was very heterogeneous and was similar to the consumer population preference distribution found by Lawlor et al. (2003). Nevertheless, after this first investigation, it is not still possible to make a relationship between the variously expressed preferences and some socioeconomic characteristics of the tested population. For instance, establishing a relationship between the ethnic group membership and the chicken preference has not been a goal of this first experiment. In addition, because the tested population was composed of citizens from a large city, Conakry, it is unlikely that this characteristic would be discriminatory. Today, people living in large African cities are less influenced in their food preference by their ethnic group membership than people living in rural African countries.

2289

2290

Sow and Grongnet

as smooth, astringent, fibrous, medicine smell, springy, brown color, white color, and chewy meat. Consumers in cluster 2 (r = 0.83) and cluster 4 (r = 0.44) expressed a high preference for the live village chicken (A) and were influenced by the chicken sensory attributes of yellow color, sweet meat, oily meat, tender, hard, and mouth persistent.

Conclusion and Implication

Figure 3. Partial least squares regression model 1 results showing the relationship between chicken sensory attributes and consumer’s overall chicken sample preference (Pref) score. Br = the intensity of the brown color of the meat; Wh = the intensity of the white color of the meat; Ye = the intensity of the yellow color of the meat; Oi = the extent of the fat odor of the meat; In = the extent of the chicken odor; Ms = smell associated with chemicals; Ro = the extent of the smell of the roast; Mf = taste remaining in the mouth after chewing and swallowing the meat; Sw = intensity of sweetness taste on the tongue; Bi = intensity of the bitter taste on the tongue; As = intensity of the meat to dry the mouth; Fi = the degree of chicken meat resistance to teeth pressure associated with a dryness feeling; Te = the intensity of the tenderness of the chicken meat; Ju = the ability of chicken meat to produce juice in the mouth; Ch = ease of chewing the chicken meat after the first bite; Sm = the level of the softness of the chicken meat between teeth; Sp = the degree of elasticity of the chicken meat; Ha = the extent to which the chicken meat requires efforts when chewing; Fb = the fibrous nature of the chicken meat found when masticating. PC1 = principal component 1; PC2 = principal component 2.

Downloaded from https://academic.oup.com/ps/article-abstract/89/10/2281/1500984 by guest on 23 January 2019

The results of this study allowed us to identify the sensory characteristics and consumer preference for chicken meat in Guinea. The sensory attributes analysis showed, in general, that there were great differences between the modern chicken and traditional chicken, mainly the live village chicken. The sensory analysis showed that the traditional live village chicken was characterized by sensory attributes such as yellow color, hard meat, oily meat, and sweet, whereas the readyto-cook broiler (imported) was characterized by white color, chewy meat, intense odor, and smooth. The live spent laying hen was assessed as a chicken meat associated mainly with bitter and fibrous attributes. The

assessors were not able to distinguish clearly between the live boiler and the ready-to-cook broiler according to their sensory attributes. Partial least squares regression with a single response showed that 72% of the chicken sensory characteristics explained 83% of the chicken preference of the consumers involved in this study. Hierarchical cluster analysis identified 4 homogenous clusters of consumers and showed that the most preferred chicken sample was the live village chicken (A) and the least preferred one was the ready-to-cook broiler (D). External preference mapping carried out on the chicken samples, their sensory attributes, and consumer clusters using a PLSR with multiple responses allowed us to characterize each consumers’ cluster by the chicken sample and sensory attributes driving its preference. Our findings showed that there was not a chicken sample that was exclusively preferred from the rest of the samples and therefore highlight the existence of place for development of all chicken categories in the local market. Poultry industry producers would find these results useful for managing the domestic chicken supply chain.

SENSORY CHARACTERISTICS AND CONSUMER PREFERENCE FOR CHICKEN MEAT

2291

ACKNOWLEDGMENTS This study would not have been possible without the support of the Guinean government, to whom we express our gratitude. We also thank the panelists and consumers who participated in this study. Our thanks also go to members of the Aid for Youth NGO for their assistance in this study.

REFERENCES Boughter, J. D., and A. A. Bachmanov. 2008. Genetics and evolution of taste. Pages 371–390 in The Senses: A Comprehensive Reference, Olfaction and Taste. G. M. Shepherd and G. Westheimer, ed. Elsevier, San Diego, CA. Carbonell, L., L. Izquierdo, I. Carbonell, and E. Costell. 2008. Segmentation of food consumers according to their correlations with sensory attributes projected on preference spaces. Food Qual. Prefer. 19:71–78. Chapman, D. M., G. Roby, S. E. Elber, J. Guinard, and M. A. Matthews. 2005. Sensory attributes of cabernet sauvignon wines made from vines with different water status. Aust. J. Grape Wine Res. 11:339–347. Delwiche, J., and M. O’Mahony. 1996. Changes in secreted salivary sodium are sufficient to alter salt taste sensitivity: Use of signal

detection measures with continuous monitoring of the oral environment. Physiol. Behav. 59:605–611. De Marchi, M., M. Cassandro, E. Lunardi, G. Baldan, and P. B. Siegel. 2005. Carcass characteristics and qualitative meat traits of the Padovana breed of chicken. Int. J. Poult. Sci. 4:233–238. Fanatico, A. C., P. B. Pillai, J. L. Emmert, E. E. Gbur, J. F. Meullenet, and C. M. Owens. 2007. Sensory attributes of slow- and fast-growing chicken genotypes raised indoors or with outdoor access. Poult. Sci. 86:2441–2449. Gallardo-Escamilla, F. J., A. L. Kelly, and C. M. Delahunty. 2005. Sensory characteristics and related volatile flavor compound profiles of different types of whey. J. Dairy Sci. 88:2689–2699. Gou, P., L. Guerrero, and A. Romero. 1998. The effect of panel selection and training on external preference mapping using a low number of samples. Food Sci. Technol. Int. 4:85–90. Hashim, I. B., A. V. A. Resurreccion, and K. H. Mcwaiters. 1995. Descriptive sensory analysis of irradiated frozen or refrigerated chicken. J. Food Sci. 60:664–666. Hough, G., and R. Sanchez. 1998. Descriptive analysis and external preference mapping of powdered chocolate milk. Food Qual. Prefer. 9:197–204. ISO. 1993. Sensory Analysis—General Guideline for the Selection, Training and Monitoring of Assessors. Part 1—Selected Assessors. ISO 8586-1 Standard. 1st ed. International Organization for Standardization, Geneva, Switzerland. ISO. 2005. Sensory analysis—Methodology—General Guidance. ISO 6658-2 Standard. 2nd ed. International Organization for Standardization, Geneva, Switzerland.

Downloaded from https://academic.oup.com/ps/article-abstract/89/10/2281/1500984 by guest on 23 January 2019

Figure 4. Partial least squares regression model 2 results showing the external preference mapping of the chicken samples, their sensory attributes, and consumer clusters. A = chicken raised traditionally by households living in villages; B = chicken raised in modern local farms and sold alive; C = chicken raised in local farms for egg production and sold after laying period; D = chicken raised in local farms for chicken meat and sold processed; E = chicken imported from other countries. Wh = the intensity of the white color of the meat; Ye = the intensity of the yellow color of the meat; Oi = the extent of the fat odor of the meat; In = the extent of the chicken odor; Ms = smell associated with chemicals; Ro = the extent of the smell of the roast; Mf = taste remaining in the mouth after chewing and swallowing the meat; Sw = intensity of sweetness taste on the tongue; Bi = intensity of the bitter taste on the tongue; As = intensity of the meat to dry the mouth; Fi = the degree of chicken meat resistance to teeth pressure associated with a dryness feeling; Te = the intensity of the tenderness of the chicken meat; Ju = the ability of chicken meat to produce juice in the mouth; Ch = ease of chewing the chicken meat after the first bite; Sm = the level of the softness of the chicken meat between teeth; Sp = the degree of elasticity of the chicken meat; Ha = the extent to which the chicken meat requires efforts when chewing; Fb = the fibrous nature of the chicken meat found when masticating. Cluster 1 = the smallest segment of consumers with 15% of the consumer sample; cluster 2 = the largest cluster, containing 36% of the chicken sample; cluster 3 = 17% of the consumer sample; cluster 4 = the second largest cluster, containing 33% of the consumer sample. PC1 = principal component 1; PC2 = principal component 2.

2292

Sow and Grongnet Northcutt, J. K., M. E. Berrang, J. A. Dickens, D. L. Fletcher, and N. A. Cox. 2003. Effect of broiler age, feed withdrawal, and transportation on levels of coliforms, Campylobacter, Escherichia coli and Salmonella on carcasses before and after immersion chilling. Poult. Sci. 82:169–173. Owens, C. M., and A. R. Sams. 1998. Meat quality of broiler breast meat following post-mortem electrical stimulation at the neck. Poult. Sci. 77:1451–1454. Ponte, P. I., I. Mendes, M. Quaresma, M. N. Aguiar, J. P. Lemos, L. M. Ferreira, M. A. Soares, C. M. Alfaia, J. A. Prates, and C. M. Fontes. 2004. Cholesterol levels and sensory characteristics of meat from broilers consuming moderate to high levels of alfalfa. Poult. Sci. 83:810–814. R Development Core Team. 2008. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Saha, A., A. V. Perumalla, Y. Lee, J. F. Meullenet, and C. M. Owens. 2009. Tenderness, moistness, and flavor of pre- and postrigor marinated broiler breast fillets evaluated by consumer sensory panel. Poult. Sci. 88:1250–1256. Santa Cruz, M. J., M. C. Martinez, and G. Hough. 2002. Descriptive analysis, consumer clusters and preference mapping of commercial mayonnaise in Argentina. J. Sens. Stud. 17:309–325. Tan, S. S., A. Aminah, Y. M. S. Affandi, O. Atil, and A. S. Babji. 2001. Chemical, physical and sensory properties of chicken frankfurters substituted with palm fats. Int. J. Food Sci. Nutr. 52:91–98. Thompson, J. L., M. A. Drake, K. Lopetcharat, and M. D. Yates. 2004. Preference mapping of commercial chocolate milks. J. Food Sci. 69:406–413. Vigneau, E., E. M. Qannari, P. H. Punter, and S. Knoops. 2001. Segmentation of a panel of consumers using clustering of variables around latent directions of preference. Food Qual. Prefer. 12:359–363. Wise, P. M., N. Bien, and C. J. Wysocki. 2008. Two rapid odor threshold methods compared to a modified method of constant stimuli. Chemosens. Percept. 1:16–23. Young, N. D., M. Drake, K. Lopetcharat, and M. R. Mcdaniel. 2004. Preference mapping of cheddar cheese with varying maturity levels. J. Dairy Sci. 87:11–19.

Downloaded from https://academic.oup.com/ps/article-abstract/89/10/2281/1500984 by guest on 23 January 2019

ISO. 2007. Sensory analysis—General Guidance for the Design of Test Rooms. ISO 8589-2 Standard. 2nd ed. International Organization for Standardization, Geneva, Switzerland. Jaeger, S. R., K. L. Rossitier, W. V. Wismer, and F. R. Harker. 2003. Consumer-driven product development in the kiwifruit industry. Food Qual. Prefer. 14:187–198. Jaturasitha, S., T. Srikanchai, M. Kreuzer, and M. Wicke. 2008. Differences in carcass and meat characteristics between chicken indigenous to northern Thailand (Black-boned and Thai native) and imported extensive breeds (Bresse and Rhode Island Red). Poult. Sci. 87:160–169. Lawlor, J. B., E. M. Sheehan, C. M. Delahunty, J. P. Kerry, and P. A. Morrissey. 2003. Sensory characteristics and consumer preference for cooked chicken breasts from organic, corn-fed, free-range and conventionally reared animals. Int. J. Poult. Sci. 2:409–416. Leclerq, B., and J. Simon. 1982. Selecting broilers for low or high abdominal fat: observations on the hens during the breeding period. Ann. Zootech. 31:161–170. Lee, Y. S., C. M. Owens, and J. F. Meullenet. 2009. Changes in tenderness, color, and water holding capacity of broiler breast meat during postdeboning aging. J. Food Sci. 74:449–454. Liu, Y., B. G. Lyon, W. R. Windham, C. E. Lyon, and E. M. Savage. 2004. Principal component analysis of physical, color and sensory characteristics of chicken breasts deboned at two, four, six and twenty-four hours postmortem. Poult. Sci. 83:101–108. McEwan, J. A. 1996. Preference mapping for product optimization. Pages 71–100 in Multivariate Analysis of Data in Sensory Science. T. Næs and E. Risvik, ed. Elsevier, Amsterdam, the Netherlands. McEwan, J. A. 1998. Cluster analysis and preference mapping. Review no. 12. Project 29742. Campden & Chorleywood Food Research Association, Chipping Campden, UK. Murray, J. M., and C. M. Delahunty. 2000. Mapping consumer preference for the sensory and packaging attributes of cheddar cheese. Food Qual. Prefer. 11:419–435. Nantachai, K., N. Hongphan, and S. Tamangklang. 2007. Sensory characteristics and preference mapping of breast and thigh chicken meat from commercial broiler and crossbred chickens. Pages 58–69 in Summer Program in Sensory Evaluation 2007. Vietnam National University, Hochiminh City, Vietnam.