Info

"Explanation of ratings: Panelists rated each product on a variety of directional questions (0 = much too little, 50 = on target, 100 = much too much).

Data for directional (DIR) questions shown, after 50 was subtracted from each rating. Positive numbers = over-delivery on the attribute. Negative numbers = under-delivery on the attribute. Like = liking rating (0 = hate, 100 = love). 'Base size: 112.

"Explanation of ratings: Panelists rated each product on a variety of directional questions (0 = much too little, 50 = on target, 100 = much too much).

Data for directional (DIR) questions shown, after 50 was subtracted from each rating. Positive numbers = over-delivery on the attribute. Negative numbers = under-delivery on the attribute. Like = liking rating (0 = hate, 100 = love). 'Base size: 112.

of systematically varied alternatives has educated the researcher and provided concrete direction for product modification.

A Case History Using Sensory Analysis—Tomato Sauce

The easiest way to understand experimental design and its link with sensory analysis is with a case history. This case history concerns tomato sauce. The issue was to de velop a new formulation to take advantage of a new processing technology.

Phase 1. Formula Selection and Design

Phase 1 comprises the selection of appropriate variables, and the recommended array of formula combinations. E&D selected 6 variables to investigate. Different experi mental designs can be used to investigate the 6 variables, depending upon the expected relation between formula variables and consumer ratings. If one expects the attributes rated by consumers to change only linearly with ingredients, with no interaction among the variables, as few as 12 prototypes need to be developed using the Plackett Burman screening design. Table 13 (part A) shows this design (31). If one expects curvature in the data, ie, the attributes (whether sensory or liking) do not follow a simple straight line versus physical changes, then one may wish to fit a quadratic curve, necessitating a quadratic design. If one expects curvature and interactions among the variables than the central composite screening design shown in Table 13 (part B) is appropriate (32). When there are simply too many prototypes to create, but one wishes to use the central composite design (because it captures curvature in the response surface and allows for interaction between pairs of ingredients) one would select the reduced design such as that shown in Table 13 (part C). R&D selected the design, shown in Table 10 (part C), encompassing 29 prototypes.

Along with the experimentally designed products, researchers test in-market competitors which provide "anchors" for the category. Panelists do not know which samples belong to the experimental design and which are actual in-markets products. All products are tested "blind," and in a randomized order. The only clues come from the panelists' sensory perceptions.

Attribute Selection for the Questionnaire

Attributes define the characteristics to which consumers will attend when evaluating the sample. It is important to select appropriate attributes, but the choice of attributes in an experimental design is not as critical as would be the case were the products to simply represent unconnected "rifle shots."

The key attribute is usually the overall criterion of acceptance, whether this be overall liking, purchase intent, or some integrative measure such as "high quality." The key attribute measures the single response that is of real interest to the investigator, and usually corresponds to the characteristic of the product that is being optimized.

The questionnaire should also deal with relevant sensory characteristics. Optimization methods can uncover formulations which are highly acceptable and simultaneously possess specific, required sensory characteristics (eg, to support an advertising position). The questionnaire can encompass image attributes as well (eg, "refreshing" for a beverage, "caloric" for a snack, etc). Image attributes are more complex cognitive characteristics.

Often there are other data available for the experimentally varied products. Additional data may include objective physical measurements, cost of goods, storage stability, ratings by experts which describe qualitative nuances, and evaluations by a quality control panel. If available, these other data sets can be added to the data file. The augmented attribute set appears in Table 14.

In this study data was available from all of these sources. Generally, however, studies involving experimental design and optimization are of a more modest scale. The

Table 13. Plackett Burman Screen Design

Product

Natural Weight Loss

Natural Weight Loss

I already know two things about you. You are an intelligent person who has a weighty problem. I know that you are intelligent because you are seeking help to solve your problem and that is always the second step to solving a problem. The first one is acknowledging that there is, in fact, a problem that needs to be solved.

Get My Free Ebook


Post a comment