Consumer acceptance of new strawberry vinegars by preference mapping

ABSTRACT

Consumers’ acceptance of new strawberry vinegars was evaluated, and trained panellists described their sensory profiles. Four strawberry vinegars, three produced from puree and one from cooked must, were evaluated jointly. Due to obtaining the highest percentage of consumer acceptance, cooked strawberry must vinegar was considered to be the best. This vinegar stood out in general impression, raisin, and liqueur attributes. Internal preference maps confirmed a higher acceptance level for the strawberry vinegars over commercial vinegar. External preference mapping, obtained by PLS2 analysis, revealed that the main sensory attributes driving consumers’ preferences are raisin, toasted caramel, spicy, and liqueur aroma.

KEYWORDS:

• Internal preference mapping
• External preference mapping
• Strawberry vinegar
• Descriptive sensory analysis
• Acceptability

Introduction

In order to survive in today’s highly competitive market, food companies need to develop new products. When these new commodities are going to be marketed, prior knowledge of consumer’s opinion is of great importance. One way to obtain such knowledge is by allowing consumers to evaluate new products by hedonic sensory test. Furthermore, the sensory characterization of products using trained judges to determine the food quality properties is also essential in the development and innovation of products.

Preference mapping is a useful tool for analysing and gaining a deeper understanding of consumer’s preferences. Indeed, in some cases, it enables to know which descriptive sensory attributes drive these preferences. Preference mapping is a graphic display created by multivariate analysis methods which enables researchers to understand how attributes influence on consumer liking,^[1,2] the differences among products,^[3,4] and products and consumers segments.^[5–7] Preference maps are also used in predicting new prototypes for industries.

There are different types of preference mapping. Internal preference mapping, for instance, uses only consumer data to determine consumers’ preference patterns. External preference mapping, however, relates consumers’ acceptability data to descriptive sensory information and/or instrumental data.^[8] Hence, the consumer liking can not only relate to product sensory attributes but also with physicochemical data.^[9]

Internal preferences mapping is a principal component analysis (PCA) of the hedonic scores matrix of across products and consumers. It is performed on a covariance matrix to allow differences to be expressed in terms of the strength of consumers’ preferences.^[10] Internal preference mapping, therefore, provides a summary of the main preference directions and the associated consumer segments.^[11] External preference mapping regresses the individual consumers’ preferences onto a set of descriptive or other analytical ratings across products.^[10] The procedure requires an objective characterisation of product sensory attributes, which is achieved by descriptive analysis, and then related to product preference ratings obtained from a representative sample of consumers.^[12]

Internal and external preference mapping techniques have been implemented in the study of a wide variety of products such as beer,^[13] tomatoes,^[2] cheeses,^[7,14] chocolate milks,^[15] filled chocolates,^[16] hot dogs,^[17] soybean paste,^[18] soy-coffee beverages,^[3] tropical juice,^[19] and cookies enriched with medicinal herbs.^[20] The objectives of the present work were: (1) to assess the sensory acceptance of four new strawberry vinegars and also to obtain their sensory profile; (2) to determine which sensory properties drive consumer liking for these products by external preference mapping techniques. Furthermore, we researched consumers’ acceptability data using internal preference mapping.

Materials and methods

Samples

The vinegars analysed in this study were produced in the laboratory of the Department of Biochemistry and Biotechnology (Faculty of Enology, University Rovira i Virgili, Tarragona) using second-quality strawberries, var. Camarosa. The strawberries were first crushed with a beater and then alcoholic fermentations were performed inoculating with a selected Saccharomyces cerevisiae yeast strain.^[21] The wine obtained was dispensed into three different containers: a glass vessel, and oak and cherry wood barrels. Each was filled with 5.5 L of wine, after which, the wine was inoculated with a selected strain of acetic acid bacteria. Thus, three different vinegars were obtained depending on container type used: FVG (glass), FVC (cherry wood barrel), and FVO (oak wood barrel). The products remained inside the barrels for a total of 2 months, the duration of the acetous fermentation process.

Furthermore, part of the strawberry puree was concentrated by heating; the resulting product was a cooked must. One litre of this must was submitted to alcoholic and acetous fermentation in the same way as non-concentrate must. Both fermentation processes took place in glass containers. The vinegar obtained was named FVCM. In addition, one common commercial wine vinegar (CWV) was used in the consumers’ acceptability test to compare the acceptability of strawberry vinegars with one normally used by consumers. This commercial vinegar was selected by expert tasters’ sensory evaluation from several common wine vinegars in the market.

Descriptive sensory analysis

An expert panel evaluated the aroma of the four samples of strawberry vinegar by quantitative descriptive analysis (QDA). QDA was performed following this method’s basic steps: development of a lexicon, selection and training of panellists, and, individual data collection of repeated measurements.^[22] Sensory descriptors were developed during preliminary tasting sessions by the panellists. First, each member was asked to describe samples with as many spontaneous descriptive terms as they found applicable.^[22] Then, finally, a list of lexicons for evaluating the strawberry vinegars was refined after consensus discussion among the panel members.^[23] The list of descriptors consisted of 16 sensory terms: pungent, ethyl acetate, herbaceous, red fruit, tropical fruit, ripe fruit, raisin, sweet, cheese/rancid, leather/animal, white wine, toasted caramel, spicy, liqueur, aromatic complexity, and general impression. The definitions used for these attributes are specified in Table 1. The selected attributes were put on a tasting-card and panellists were asked to rank each descriptor on a 10 cm unstructured scale (from unnoticeable to very strong). The panel was composed of eight tasters (six women and two men). All of the panel members, belonging to the laboratory staff, were trained according to international protocols.^[25] Training was performed over 2 months, two sessions per week, and was based on ranking solutions prepared at different concentrations with reference substances of the trained attribute (Table 1). All solutions tested by the panel had a final acetic degree of 6% v/v acetic acid, except for pungent descriptor. With this training, the judges learned the odour of selected attributes and also their intensity. Panellists were trained in aromatic complexity by increasing the complexity of tasting solutions through the progressive addition of substances to the mixture tested. For general impression attribute, solutions with a mixture of different reference substances, some with a pleasant odour and others with an unpleasant odour, were used. Fifteen millilitres of vinegar sample or reference solutions were presented in a dark glass covered with a plastic dish in the tasting or training sessions, respectively. The glasses were identified with three-digit random codes. The different samples were placed randomly in each test. The panel’s performance was evaluated using repeatability and reproducibility measures.^[26,27]

Table 1. Definitions of the attributes used by the expert panel to describe the samples.

Descriptors	Definitions	References
Pungent	Sensation produced by an aqueous solution of acetic acid^a	Acetic acid solution^a
Ethyl acetate	Odours associated with glue and nail polish remover^a	Ethyl acetate solution^a
Herbaceous	Odours associated with green leaf or cut grass	cis-3-Hexen-1-ol
Red fruit	Odours associated with strawberry or raspberry^a	Mixture of juice strawberry and raspberry
Tropical fruit	Odours associated with peach, apricot, mango, or passion fruit	Tropical fruit juice
Ripe fruit	Odours associated with overripe fruits	Mixture of several overripe fruits (banana, apple, peach)
Raisin	Odours associated with grape raisin^a	Juice of grape raisin^a
Sweet	Odours associated with cotton candy or vanilla	Cotton candy solution
Cheese/rancid	Odours associated with aged cheese^a	Isovaleric acid solution
Leather/animal	Odours associated with leather^a	4-Ethylphenol solution
White wine	Odours associated with young white wine	Young white wine^a
Toasted caramel	Odours associated with toasted sugar	Toasted sugar
Spicy	Odours associated with clove or cinnamon	Maceration of clove and cinnamon
Liqueur	Odours associated with liquorice, brandy, or aged Sherry wine	Sotolon solution
Aromatic complexity	Wide variety of aromatic nuance	Solution with different number of reference substances mentioned above
General impression	Aromatic overall impression originated by the volatiles released during the sniffing of the sample	Mixture of substances with pleasant and unpleasant odours

^aDescriptor or reference taken from Tesfaye et al.^[24]

Consumer acceptability test

A total of 52 regular consumers of vinegar were recruited from the IFAPA Research Centre (Córdoba). The number of consumers consulted was limited by the sample volume available. Ages ranged from 20 to 60 years, 61% being women and 39% men. Consumers were asked to evaluate their odour and taste liking using a 9-point hedonic scale, where 1 = dislike extremely and 9 = like extremely. The four vinegars were evaluated by consumers in a single session. Fifteen millilitres of the sample were presented in coded black glasses covered with a plastic dish. First, consumers were asked to smell the samples. Then, in order to evaluate the samples’ taste as a condiment, consumers put several drops onto a piece of lettuce and tried each one.

Statistical analysis

ANOVAs were conducted to assess significant differences (p < 0.05) among consumers’ acceptance scores of the different vinegar samples and also to evaluate significant differences (p < 0.05) among attribute scores given by the expert panel in the QDA. To interpret the consumers’ preferences, a PCA based on covariance matrix of consumers (variables) by products (objects) was performed to obtain an internal preference map. Partial least square (PLS) was applied to link consumers’ preferences to sensory data. PLS2 was performed using the descriptive variables (16 attributes × 4 samples) as regression matrix (X data) and the consumers’ odour preference scores (52 variables × 4 samples) as regressand (Y data). ANOVAs and PCA were performed with Statistica, version 7.0 software (Statsoft, Tulsa, USA) and PLS2 was conducted with the statistical software Unscrambler, version 9.1 (Camo Process AS, Oslo, Norway).

Results and discussion

Descriptive sensory analysis

The aroma of the vinegars was assessed by 8 tasters, using 16 descriptors (pungent, ethyl acetate, herbaceous, red fruit, tropical fruit, ripe fruit, raisin, sweet, cheese/rancid, leather/animal, white wine, toasted caramel, spicy, liqueur, aromatic complexity, and general impression). The descriptors agreed upon previously as being the best for describing the sensory characteristics of the vinegars studied. Table 2 shows the results obtained in the descriptive analysis. As can be seen, the vinegar from cooked must and that acetified in oak barrel presented the highest scores in general impression and aromatic complexity, the first presenting significant differences. These attributes reflect the overall aromatic quality. Moreover, the FVCM sample stood out for its raisin and liqueur aroma, FVC for its unpleasant cheese aroma, and FVO for its red fruit odour, with statistically significant differences in all cases. According to the expert panel, vinegars from strawberry puree had a characteristic aroma of tropical fruit.

Table 2. Intensity of sensory attributes of studied vinegars.

	FVG	FVC	FVO	FVCM
Ethyl acetate	0.2 ± 0.5^a	0.15 ± 0.25^a	0.4 ± 0.7^a	0.2 ± 0.3^a
Pungent	3.0 ± 1.5^ac	4.1 ± 1.3^ab	4.3 ± 1.1^b	2.5 ± 0.8^c
White wine	0.3 ± 0.5^a	1.6 ± 0.6^b	0.06 ± 0.17^a	0.03 ± 0.09^a
Herbaceous	1.1 ± 1.4^ab	0.2 ± 0.5^a	0.4 ± 0.5^ab	0.9 ± 1.1^b
Red fruit	3.2 ± 1.8^a	1.10 ± 0.23^b	5.7 ± 1.4^c	3.1 ± 1.4^a
Tropical fruit	5.2 ± 2.0^a	6.4 ± 1.1^a	5.7 ± 1.7^a	0.8 ± 0.8^b
Ripe fruit	5.5 ± 1.3^a	2.6 ± 1.7^b	5.4 ± 1.0^a	5.8 ± 1.5^a
Raisin	0.5 ± 0.6^a	0.3 ± 0.6^a	2.3 ± 0.9^b	6.6 ± 0.9^c
Sweet	5.2 ± 1.6^a	3.6 ± 1.3^b	6.7 ± 0.8^c	6.4 ± 1.4^abc
Liqueur	0.2 ± 0.4^a	0.2 ± 0.6^a	0.04 ± 0.7^a	4.8 ± 0.4^b
Toasted caramel	0.2 ± 0.5^a	0.02 ± 0.08^a	0.1 ± 0.3^a	0.3 ± 0.6^a
Spicy	0.0 ± 0.0^a	0.03 ± 0.11^a	0.04 ± 0.15^a	0.5 ± 0.8^b
Cheese/rancid	0.2 ± 0.5^a	6.6 ± 0.7^b	0.4 ± 0.6^a	0.1 ± 0.3^a
Leather/animal	0.4 ± 0.7^ab	1.2 ± 1.5^a	0.4 ± 0.7^ab	0.1 ± 0.3^b
Aromatic complexity	2.9 ± 1.1^a	2.8 ± 0.8^a	4.8 ± 1.0^b	5.3 ± 0.5^b
General impression	4.4 ± 1.6^a	2.5 ± 0.8^b	5.1 ± 0.7^a	6.6 ± 1.5^c

Different letters in the same line indicate significant difference (p < 0.05).

In a previous study, red wine vinegar acetified in cherry wood barrel presented a higher score in the “red fruit” attribute than wine vinegar produced in an oak wood barrel.^[28] In our case, this was not observed. In strawberry vinegar, the main source of red fruit aroma is the raw material. For this reason, we did not observe the expected effect caused by the influence of the cherry wood barrel. In contrast, strawberry vinegar fermented in a cherry wood barrel reached lower score in the “red fruit” attribute than the strawberry vinegar produced in an oak wood barrel. This could be due to the powerful odour of cheese odour that masked the red fruit aroma in our vinegar produced in cherry wood barrel. However, with respect to the “sweet” attribute, the results found were consistent with Cerezo et al.^[28]

A PCA was performed on the data obtained in QDA. The first two components explain 89.3% of the total variance (Fig. 1). As shown in the figure, the vinegar from cooked strawberry must appears in the lower left quadrant and separate from the rest. The second principal component (PC2) also separates vinegar from strawberry puree produced in cherry wood barrel from the other strawberry puree vinegars. According to the loadings, vinegar from cherry wood barrel was correlated with cheese/rancid, leather/animal, and white wine sensory attributes (Fig. 1b); the two first being aromatic defects. This means that this sample is located in the opposite direction to the general impression attribute with regard to the PC1. Vinegar from oak wood barrel was correlated with the attributes of red fruit and ethyl acetate, the first being characteristic of this sample, as mentioned above. Finally, vinegar from cooked strawberry must was mainly correlated with spicy and liqueur attributes and also with raisin and toasted caramel. Therefore, PCA confirms that there are some differences among aromatic attributes in vinegar from strawberry puree and vinegar from cooked strawberry must.

Figure 1. Data scores (a) and variable loading (b) plots on the planes made up of the first two principal components (PC1 against PC2).

Consumers’ sensory acceptability

Table 3 presents the mean liking rating for each vinegar sample. The highest acceptance level was for cooked strawberry must vinegar. However, no significant differences (p < 0.05) were found between the developed vinegar and strawberry vinegars. These results can be explained since the intravariability of consumers’ preferences is higher than the intervariability among samples. The commercial vinegars obtained the lowest mean score with significant differences with respect to the other samples. This vinegar is produced from white wine by submerged acetous fermentation, while strawberry vinegars were produced by surface acetous fermentation. The first type of fermentation, on an industrial scale, takes place in stainless steel bioreactors, open acetification systems with forced aeration, which leads into important losses of volatile compounds.^[29] As a result, the products obtained by this method are aromatically poor. In contrast, previous results have shown that the strawberry vinegars used in this study retain the impact odorants of strawberry.^[30] Therefore, different raw materials as well as different production processes endow the vinegars with organoleptic differences. These two primary differences could be the reason for consumers scoring the commercial vinegar and strawberry vinegars differently.

Table 3. Mean consumer’s liking rating for each vinegar sample.

	Samples
	FVG	FVC	FVO	FVCM	CWV
Overall odour liking	6.2 ± 2.8	5.6 ± 2.2	6.1 ± 2.6	6.6 ± 2.7	3.9^a ± 2.2
Overall taste liking	6.1 ± 2.5	5.6 ± 2.3	6.4 ± 2.7	6.44 ± 2.5	5.1^b ± 2.3
Overall acceptance^c	6.2 ± 2.6	5.6 ± 2.2	6.2 ± 2.6	6.50 ± 2.6	4.6^a ± 2.3

^aSignificant difference (p > 0.05) with all other samples.

^bSignificant difference with other samples except FVC.

^cThis value was a mean between odour and taste liking values.

In Fig. 2, the percentages of consumers who chose each vinegar in first, second, third, fourth, and fifth place by preference are given. With respect to odour liking (Fig. 2a), the percentages showed that the sample chosen as first by the highest percentage of consumers was the cooked must vinegar matching with the sample which obtains the highest preference score. Furthermore, a high percentage of consumers gave the second position to the vinegar produced in a glass container, the third and fourth for vinegar produced in oak and cherry barrels, respectively. Half of consumers liked the commercial vinegar least.

Figure 2. Percentage of consumers who chose each vinegar in descending order of preference (a: overall odour liking; b: overall taste liking).

Except for commercial vinegar, ranked fourth by a large amount of consumers, the order of vinegar with regard to the consumers’ taste preference was not so clear (Fig. 2b). However, we can say that the cooked must vinegar was again preferred by consumers, since a high percentage of them classified it as first or second. On the other hand, it seems that vinegar produced in cherry barrel obtained a slightly higher score in taste than those produced in a glass container or in oak barrel. Moreover, consumers considered that the vinegar produced in cherry barrel presented better taste than odour.

Internal preference mapping

The internal preference maps (Figs. 3 and 4) were created by the PCA of a data matrix with consumers (variables) and products (objects). With regard to the odour of vinegars, PCA results showed that the first two principal components explained 66.67% of total variance among acceptation of samples. Figure 3a shows scores of samples plotted onto the plan of these two principal components. We can observe the samples separate into three groups: one formed by commercial vinegar, another by strawberry vinegar produced in a glass container, and the third comprising strawberry vinegars produced in wooden barrels, as well as the vinegar produced from cooked must. These two last sample groups appear closer in the graph, separated from the other vinegars by PC2. In Fig. 3b, each point represents the correlation between the consumers’ acceptance data and the first two components. Therefore, most consumers are placed in the right quadrants, being positively correlated with the first component, consequently preferring the vinegars obtained from strawberry puree. Although FVCM reached the highest mean acceptability score, the number of consumers correlated with this sample was not higher than with the other strawberry vinegars. However, if we consider PC2 (Fig. 3), there was a large number of consumers located in the two upper quadrants also showing a high acceptance of the vinegars located in said quadrants (FVC, FVO, and FVCM). The low number of consumers located in the same quadrant on the graph as CWV confirmed the consumers’ low acceptability of this vinegar.

Figure 3. Internal preference map obtained by PCA on the consumers’ overall odour preference data: (a) sample score plot; (b) consumers’ loading plot.

Figure 4. Internal preference map obtained by PCA on the consumers’ overall taste preference data: (a) sample score plot; (b) consumers’ loading plot.

In relation to the sample taste, the first two PCs explained 63.35% of total variance. In this case, the graph of sample scores into the plan of the first two principal components revealed that the consumers’ preferences for the vinegars differ greatly, and that the samples are widely separated except for FVCM and FVC (Fig. 4a). With regard to the loadings consumers’ acceptance (Fig. 4b), there is a greater number of consumers negatively correlated with the first component, showing a preference for the samples mentioned above. In the case of CWV, the low consumer acceptance value was confirmed by the internal preference map since there were very few consumers linked to this vinegar.

External preference mapping

After analysing of consumers’ preferences, we were interested to know which aromatic characteristics determined this consumer preference pattern. The data of consumer preferences and sensory descriptive analysis data of the four strawberry vinegars were used to construct the external preference map (Fig. 5). PLS2 analysis conducted for external preference mapping explained 94% of sensory variance and 55% of liking variance. Most consumers are correlated negatively with the PC1, and located in the same direction as the FVCM sample.

Figure 5. External preference map obtained by PLS2 analysis of descriptive data and consumers’ overall odour liking of strawberry vinegars: (a) sample score plot; (b) correlation loadings of sensory descriptive and consumers’ overall odour liking data.

These results are consistent with the earlier ones; FVCM reached the highest mean score of overall odour liking. In order of preference, it was chosen as first by the highest percentage of consumers. The attributes with a high positive relation to preference were raisin, toasted caramel, general impression, spicy, liqueur, aromatic complexity, sweet, ripe fruit, herbaceous, red fruit, and, to a lesser extent, ethyl acetate. Therefore, projecting the sensory attributes onto external preference mapping enabled raisin, toasted caramel, spicy, and liqueur aroma to be identified as the major drivers of acceptance. Toasted caramel is the aroma character of two important volatile compounds found in strawberry: furaneol and mesifurane. Due to their low odour threshold and their high quantities in strawberry, these two compounds are considered as strawberry impact odorant.^[31] Attributes related to unpleasant aromatic notes such as cheese/rancid, leather/animal, or pungent, among others, were positioned to the right side of the map, where there were fewer consumers. As expected, these attributes had a negative correlation with preference.

Conclusion

In consumer acceptability test, strawberry vinegars reached higher scores than a commercial white wine vinegar. This fact indicates that strawberry vinegars could be commodities with a good market acceptance. Descriptive sensory analysis has enabled us to conclude that during the strawberry vinegars’ production processes, the cooking of the substrate and the use of oak wood barrels result in vinegars with a higher score in pleasant aromatic nuances. Internal preference mapping provided us with rapid data interpretation. Internal preference mapping revealed that there were differences in the odour and taste preferences of consumers with regard to the vinegars studied. In terms of odour, vinegars from unconcentrated strawberry puree were preferred by more consumers, while cooked must vinegar and that produced in cherry wood barrel were considered to be the best by a high number of consumers for their taste. These results showed that consumers preferred strawberry vinegars over a commercial white wine vinegar. External preference mapping, where only strawberry vinegars were considered, showed that the attributes that mainly drive consumers’ preferences were raisin, toasted caramel, spicy, and liqueur aroma.

Acknowledgements

The authors thank Agromedina enterprise for providing the fruit substrates and Dr. A. Mas’ research group from University Rovira i Virgili for providing fruit vinegars. We also thank Dr. Francisco Peña for his invaluable help in performing the consumer’s preference tests.

Funding

This research was made possible through the financial support from the Spanish government by means of a predoctoral grant and the research project AGL2007-66417-C02-01 funded by the Ministry of Science and Innovation.

Additional information

Funding

This research was made possible through the financial support from the Spanish government by means of a predoctoral grant and the research project AGL2007-66417-C02-01 funded by the Ministry of Science and Innovation.