Printed matter – Method
Reexamination Certificate
1997-09-03
2003-09-23
Wellington, A. L. (Department: 3722)
Printed matter
Method
C283S056000, C283S070000, C434S365000
Reexamination Certificate
active
06623040
ABSTRACT:
FIELD OF THE INVENTION
The invention relates to consumer testing in general and, more particularly, to consumer preference testing.
BACKGROUND OF THE INVENTION
In consumer testing, e.g., testing that a company conducts to determine how well its product will be received by the intended consumer group, there are a number of different testing protocols that are followed. In hedonic testing, a consumer is presented with product samples, e.g., salty snacks such as pretzels or potato chips; cookies; toilet paper; etc., and is asked to evaluate each of several different products of a given type on a number of different qualities or product characteristics. For example, the test subject may be asked to evaluate potato chips for saltiness, flavor, appearance, texture, “mouth feel,” etc.
When conducting hedonic testing, it is customary to have the test subject rate each parameter on a Likert scale, i.e., a scale having discrete integer values ranging from a low value (e.g., disliked extremely) to a high value (e.g., liked extremely). Likert scales for hedonic testing typically require a response from 1 to 5, 1 to 7, or 1 to 9. For consumer testing of food products, a nine-point Likert scale customarily is used, with a response of 9 indicating that the consumer liked the particular quality being evaluated extremely; a response of 8 indicating that the consumer liked the quality being evaluated very much; 7 indicating that the consumer liked it moderately; 6 indicating that the consumer liked the quality slightly; 5 indicating that the consumer neither liked nor disliked the quality; 4 indicating that the consumer disliked the quality slightly; 3 indicating that the consumer disliked it moderately; 2 indicating that the consumer disliked it very much; and 1 indicating that the consumer disliked the quality being evaluated extremely. In addition to evaluating the individual qualities, the test subject typically is asked to provide a Likert scale rating indicating how well he or she liked the product overall, i.e., taking into account all factors or qualities which he or she has been or will be asked to evaluate.
Usually, the test subject is asked to indicate his or her liking or evaluation of the same characteristics or qualities of several products of the same type, e.g., several brands of pretzels or cookies. In sequential monadic testing, the consumer provides a complete evaluation for a given product before providing a complete evaluation for another product. In other words, a consumer testing potato chips, for example, indicates his or her liking of the taste, crunch, saltiness, aroma, and overall liking of brand X potato chips, then indicates his or her liking of the taste, crunch, saltiness, aroma, and overall liking of brand Y potato chips. (This is in contrast to having a consumer provide a response for, e.g., the taste of brand X, then the taste of brand Y, then the crunch of brand X, then the crunch of brand Y, then the saltiness of brand X, then the saltiness of brand Y, etc.)
Finally, when the consumer has completed his sequential monadic evaluation of all the products being tested—typically no more than three or four different products are tested in a given test setting—the consumer is asked to indicate which of two given products he or she likes better, e.g., whether he likes W or X better; whether he likes W or Y better; whether likes W or Z better; whether he likes X or Y better; whether he likes X or Z better; and whether he likes Y or Z better. This latter testing is referred to as forced choice testing, i.e., the consumer is forced to choose which product he or she prefers from among a forced selection of two possible products.
In certain situations, it is not possible or feasible to conduct forced choice testing, e.g., when the product flavors are so strong that tasting more than two products, or tasting the products a second time to select a preference, is precluded. Alternatively, time limitations may be such that adding the task of additional forced-choice comparisons among all combinations of products is unreasonable and/or excessive. Thus, in such situations, it may be difficult to obtain meaningful forced choice data. For example, as the number of products being evaluated increases, so too does the number of possible forced choice pairings, but at a faster rate than the number of products increases. Other difficulties or problems with forced choice testing include the following: 1) the most accurate and reliable product assessment generally is considered to be the initial reaction to the product, not a response that comes after answering questions about individual taste, appearance, and texture details, etc.; 2) the respondents' decreasing memory of the samples becomes problematic if the samples are not available for re-sampling, but sensory fatigue (overload of the taste senses) becomes an issue if they are; and 3) the last product seen and sampled and the last question asked may influence the preference unevenly.
SUMMARY OF THE INVENTION
In the present invention, we have recognized and taken advantage of the fact that there is a correlation between a consumer's overall like or dislike of two particular products (hedonic test results), each evaluated individually, and the likelihood that the consumer will choose one product over the other (forced choice test results) in a forced choice comparison. Thus, it is possible to determine the likelihood that a consumer will choose one product (product A) over another product (product B) of the same type simply by conducting hedonic testing and without specifically conducting forced choice testing.
Thus, according to a first aspect of the invention, a method is provided for determining forced-choice preference information from hedonic test data. The method entails obtaining a test group of human test subjects and providing each of the test subjects in the group with two or more test samples. The test subjects are asked to sample each of the two or more test samples and provide hedonic test data indicative of how well they liked each of the test samples. The hedonic test data is then processed from all of the test subjects to determine at least one predicted forced choice preference result, the predicted forced choice preference result being indicative of the likelihood that a test subject would select one of the test samples over another one of the test samples in a forced choice comparison between a selected pair of test samples.
In a preferred embodiment of the method of the invention, of a pair of test samples being compared, a stronger test sample and a weaker test sample are determined. For each test subject, an overall-liking difference value is determined, the overall-liking difference value being based on the overall-liking value the given test subject assigned to the stronger test sample minus the overall-liking value the given test subject assigned to the weaker test sample. Then, for each possible overall-liking difference value, the number of occurrences of that possible value is determined. For each possible overall-liking difference value, the number of occurrences of that value is multiplied by a probability coefficient to determine a corresponding subset number which represents the number of test subjects in a correlative, proportionate subset of the test subjects yielding that particular possible overall-liking difference value that would choose the stronger test sample over the weaker test sample in a forced choice comparison between the two. The subset numbers corresponding to each of the possible overall-liking difference values are then totaled to determine the total number of test subjects who would choose the stronger test sample over the weaker test sample in a forced choice comparison between the two. The preference for the stronger test sample may then be expressed by dividing the total number of subjects who would choose the stronger test sample over the weaker test sample in a forced choice comparison by the number of test subjects in the whole test group.
Preferably, the stronger t
Foley Mary M.
Lancaster Beaufort M.
Henderson Mark T.
Recot Inc.
Rothwell Figg Ernst & Manbeck
Wellington A. L.
LandOfFree
Method for determining forced choice consumer preferences by... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method for determining forced choice consumer preferences by..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method for determining forced choice consumer preferences by... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3001586