:: Sign up for our Newsletter
* Name

Your name is required.
* Email

Your email is required.
* Mandatory

Discriminant Analysis / Logistic Regression Analysis / Logit Analysis

Questions that the analysis can address:

  1. Which customers are most likely to purchase a product or subscribe to a service?
  2. What characteristics are most important in driving preferences or predicting outcomes?

URIKA Research employs these predictive techniques in cases where the goal is to predict group membership from a set of characteristics or variables. For example, we may want to learn which variables predict preferences between two competing brands.

In Discriminant analysis, we determine the extent to which the set of predictors (which must be continuous as opposed to discrete variables) can be combined into a "function", which is then used to predict which group a respondent belongs to. If the function is accurate in predicting actual brand preference (the function predicted that the respondent would prefer Brand X, and the respondent actually does prefer Brand X), we learn which combination of characteristics is most important for "discriminating" between preferences for Brand X and preferences for Brand Y.

We use Logistic Regression analysis to predict group membership. However, unlike Discriminant analysis, the predictors can be either continuous (e.g., a score on an index) or discrete (gender, occupation).

We also use Logit Analysis when the goal is to predict a discrete outcome (e.g., occupational category) from a set of discrete predictors.