Discriminant Analysis

Main aspects

  • Identifies those characteristics which best predict membership of certain groups
  • In a sense, the reverse of cluster/segmentation analysis - starts with the groups and finds the best predictors
  • For example, can identify which statements concerning taking medication might best predict whether someone is likely to be compliant or not
  • Knowing a respondent's answers to perhaps just a handful of questions can fairly reliably inform us of another of their characteristics
  • Can subsequently be used to quickly identify certain 'types'