This paper introduces a statistical algorithm developed to avoid conscious rationalization bias and to better predict consumers' preferences for food products. The approach taps into consumer's mind implicitly and yet explicitly as in traditional surveys.
We assumed that conclusiveness of binary responses is associated with the proportional differences in reaction time in the forced choice questions. The reaction time, arguably the intuitiveness, is used as a weight function to compensate consumer responses. Responses with relatively long reaction time are likely less intuitive and likely less conclusive. The potential effect of tiredness/fatigue is considered.
The utility of an algorithm is demonstrated using data from a marketing survey of two competing products. Each of 300 participants performed 30 preference tests between a pair of attributes randomly selected among 45 attributes. The reaction time was recorded and used in the multivariate analysis with attributes. The results provide a better picture of the differences between the products.
|