S: Sensory & Food Quality Drivers of Liking for Soy-Based Indian-Style Extruded Snack Foods Determined by U.S. and Indian Consumers Erika A. Neely, Youngsoo Lee, and Soo-Yeun Lee Abstract: Although many researchers have studied potential ways to deliver soy in novel forms, little is known about specific sensory attributes associated with soy snacks, or how those attributes drive liking for consumers. The first objective of this study was to use sensory descriptive analysis to characterize 9 extruded soy snacks with varying soy levels and soy grits contents. A total of 12 trained panelists used a descriptive analysis method to evaluate the snacks and found 14 attributes to be significantly different across the samples. Furthermore, it is not known how preferences of Indian snack consumers living in the United States and India may vary for sensory attributes of soy snacks. The 2nd objective was to correlate descriptive profiling data and previously collected consumer data to construct preference maps illustrating consumers’ attitudes toward the snacks. Results indicate that consumers generally accept samples characterized by attributes such as crunchy, cumin, curry, salty, and umami, but dislike samples with wheat, rough, or porous attributes. Indian consumers differed from the U.S. consumers in that their preferences were more varied, and they tended to be more tolerant of wheat and porous attributes. Therefore, different strategies should be utilized when developing products for these groups to cater to their specific inclinations. Keywords: consumer acceptance, descriptive analysis, extrusion, India, soy Introduction Product developers are charged with the task of producing prod- ucts that appeal to their target consumers. Conducting consumer acceptance testing is, of course, one obvious part in determining if products meet the desires of consumers. However, it is generally recommended that consumers should not be used to determine specific sensory attributes that contribute to their liking (or dislik- ing) of a product, since they are evaluating products without any training, and therefore, are not able to precisely describe character- istics about the products (Lawless and Heymann 1999). Although consumer tests can give valuable information about the accept- ability of products, they do not give much specific guidance to the product developer on how to go about improving a product’s sensory characteristics to obtain a better product. Descriptive analysis methodology, on the other hand, is able to provide detailed depictions of products’ sensory attributes in an objective manner. It uses a trained panel of judges to evaluate the characteristics of products and their intensities so as to develop a complete sensory profile (Lawless and Heymann 1999). The judges are acting as analytical instruments, making reproducible measurements about the products. These judges, however, should not be asked to make hedonic evaluations of the products, just as consumers should not be asked to rate attribute intensities. Integration of consumer and descriptive data is imperative to determining what sensory characteristics most impact consumer MS 20090932 Submitted 9/21/2009, Accepted 4/15/2010. Authors are with Univ. of Illinois at Urbana-Champaign, Dept. of Food Science and Human Nutrition, 905 S. Goodwin Ave., Urbana, IL 61801, U.S.A. Direct inquiries to author Soo- Yeun Lee (E-mail: soolee@illinois.edu). acceptance. Attributes that are responsible for acceptance of a product are often termed “drivers of liking” (Moskowitz 2001). One way to analyze drivers of liking is by using preference map- ping methodologies (MacFie 2007). Preference maps are visual presentation of complex data that allow interpretation of indi- vidual consumer preference trends. The method utilizes principal component analysis to reduce the dimensions of variation so that a large number of consumers’ responses can be represented on just a few dimensions, often two, so as to allow presentation of a standard Cartesian coordinate graph. Construction of external preference maps involves first plotting products onto a sensory map using the attributes evaluated by a trained panel. Consumer acceptance data are then regressed onto this map to visually analyze how con- sumers relate to the products, and to determine what attributes drive consumer liking. Using this information, product developers can manipulate sensory qualities of prototype products to move them around the map to optimize consumer liking. This method has been used previously to study a variety of products, including chocolate milks (Thompson and others 2004), cheddar cheeses (Young and others 2004), sausages (Grewal and others 1974), and lager beers (Guinard and others 2001). Product development for international markets presents its own set of challenges, some of which are associated with the time and expense required to travel to the country of interest to conduct consumer testing. Therefore, there is interest in using “surro- gate consumers,” consumers who live in the same location as the product developers but could serve as a model for the true pop- ulation of interest based on criteria that make them similar to the international market segment, such as ethnicity or country of origin. In a previous study by Yeh and others (1998) compar- ing Korean, Chinese, and Thai individuals living in the United States to counterparts living in their respective native countries, C  2010 Institute of Food Technologists R  S292 Journal of Food Science  Vol. 75, Nr. 6, 2010 doi: 10.1111/j.1750-3841.2010.01672.x Further reproduction without permission is prohibited