Problems of Adequacy and Validity in Focus-Group Method and Method of Semantic Analysis

One of the greatest problems in sociological research by method of focus-groups is the fact that the interpretations submitted by members of the studied group sharing one mentality are reinterpreted by expert-researchers who belong to a different mentality. Since attitude interpretation mechanisms are subconscious, there is practically no way to avoid erroneous judgments. Besides that, identification of basic attitudes for a given mentality concerning a certain problem is carried out under a moderator's supervision. A moderator might have other motives and goals than those of the group. This inevitably leads to a bias in group reaction. Another factor that distorts the variants of interpretation and their probability within a group is a team discussion of a problem. It is well-known that associative processes within a group tend to become uniform. According to F.Allport's research, generation of original ideas is two times lower in the presence of other people. Basically, focus-group is a variant of brain-storming procedure, because it combines elements of group and individual interview concerning a problem that presents interest to a moderator. Research testifies that ideas produced jointly through a brain-storm tend to be average in terms of number and lower in quality than those generated individually. The only exception are problems of technical design, strictly formalized and well-performed both by a team and a moderator. Goals of studying subconscious attitude don't belong to teamwork type, because it is those attitudes that form different interpretations for a given mentality. Informal leaders emerging spontaneously in a team also become a bias factor in research results, because they strongly influence team-members.

Hence, the major disadvantages of a focus-group method in market research would be the following:  

1. This method is used for studying the quality estimate that determines the significance of a certain set of properties for successful market advance of a product. However, this estimate fails to reveal the precise correlation of these qualities that determine the market rating of a product.

2. Standard marketing decisions are based on subjective reactions, of which the informants are aware. There are subconscious objective attitudes that actually determine the consumer's choice of goods.

3. The process of revealing the subconscious attitudes within a focus-group on the basis of expert judgement depends on a number of subjective factors. This in itself is not valid. Expert analysis of attitudes in the context of this method is purely qualitative and, therefore, it can not serve as a quantitative criterion for designing a forecast rating for the product while it's still being manufactured.

4. Quantitative changes in standard technologies are reduced, at their best, to the probability estimation for one or another opinion. However, they either fail to reflect the quantitative expressedness of the studied characteristics, or present rough estimates thereof on a 3-7 score scale. Quantitative methods of rating estimation by attitudes (objective characteristics) are completely absent, although they give the most real forecast.

The method of Semantic Analysis is free from the above limitations.

Obviously enough, the criterion of product proximity to the ideal must determine its market rating in a consumer's mind. The proposed procedure defines the ideal not only in the framework of subjective evaluations, but also within subconscious objective attitudes of the studied mentality. Our procedure reveals the most stable attitudes of the sample studied, on which basis further calculation is made. This yields valid quantitative and qualitative results.

The computer automatically performs description of the ideal immediately after computer testing. Calculation is made on the basis of qualities selected and familiar ratings of the goods tested. The ideal product is mathematically identified according to the criterion of proximity to the ideal. The information on metric properties of the ideal makes it possible to forecast the progress of a product designed.

So, problem solving breaks into two stages. The first stage is marked by identification of the ideal parameters, while the second one involves comparison of the product designed to the ideal calculated. The second stage can multiply repeat  through the software supplied to a customer. It does not involve outside research assistance. In fact, there can exist a continuous express monitoring of a product designed.

Back