Question
what methods exist to develop a consistent measurement scale from data about pairwise comparisons of a consumer's preference among alternative product features and costs.
Mark
Answer
As part of my undergraduate education I learned about an engineering tool called a "criteria function". I will attempt to describe it, it is a table that is very well suited to spreadsheets like Excel. Label the rows on the left with features, costs, and benefits. In the second column, next to each feature rank the importance of each feature, I use a scale from 0 to 1 (0.1, 0.2, 0.5, etc). Label each additional column as "solution 1", "solution 2", "solution 3", etc.For each solution and feature, put a number in the corresponding cell that is how well the solution being considered satisfies the requirement of the feature. Again, I use a fractional scale from 0 to 1.When dealing with consumers, you can present the features only two at at a time "is feature a more important than feature b". This can be used to develop a ranking of the user's perception of the importance of each feature. An engineer would use a more quantified measure, power, size, weigjht, cost, efficiency, etc.When ranking the solutions use a similar pairwise comparison, the key is to give the interviewee only two choices at each step "Which worked better for feature 1, solution A or solution B"When you put all the rankings into the spreadsheet, multiply the importance of the feature by how well it performed for each solution, then sum these products for each solution to produce an overal performance number for that solution. The solution with the highest performance number wins. If the construction of the table isn't clear, let me know and I will send an example spreadsheet. http://vrd.askvrd.org/services/answerschema.xml
