Análisis MaxDiff

What is MaxDiff analysis

MaxDiff or Maximum Difference Scaling is a method introduced by Jordan Louvière that allows to obtain importance of attributes. Attributes are presented to a respondent and he has to choose the best and worst attributes (most important / least important)

Two steps are needed to apply that method. First, a design has to be generated so that each attribute is presented with other attributes an equal number of times. Then, once the respondent has selected for each choice the best and worst attribute, a model is applied in order to obtain importance of each attribute. A Hierarchical Bayes model is applied to obtain individual values of the importance.

In the case of MaxDiff models, individuals have to choose between selections of attributes. Thus, a number of choices are given to all individuals (we will select an attribute from a number of attributes).

Analysis of these choices is made using a hierarchical Bayes algorithm which gives individual results. Parameters are estimated at the individual level using an iterative method (Gibbs sampling) taking into account each individual’s choice but also the global distribution of the choices. The obtained individual importance will be more precise.

The MaxDiff analysis allows to obtain individual MaxDiff score for each respondent and each attribute.