What Is Manova Discriminant Analysis

Data are often graphed to visualize correlations between variables.

Statistics examines correlations between sets of data. In cases where factors, such as age, affect an outcome, for example, salary, the affected outcome is typically called the dependent variable, while the factors are called independent variables. Discriminant analysis determines the relationships between independent and dependent variables.

Discriminant Analysis vs. MANOVA

Discriminant analysis determines which factors, the independent variables, most strongly affect an outcome, the dependent variable. MANOVA, which stands for Multivariate Analysis of Variance, is like discriminant analysis with the dependent and independent variables reversed. In MANOVA, the factors are treated as dependent variables and the outcome is the independent variable.

Uses

Discriminant analysis and MANOVA can both be used to identify which factors have the strongest effect on an outcome. They can also be used to weed out irrelevant factors or to classify outcomes based on a prediction model. In addition, these statistical methods can be used to test a hypothesis about how different predictors cause an outcome.

Assumptions

Like many statistical tests, discriminant analysis and MANOVA require an adequate sample size and assume that errors are randomly distributed within the test data. Discriminant analysis also assumes that the independent variables are not correlated with each other.

References

• "Discriminant Function Analysis (DA)"; John Poulsen, Aaron French:
• "Discriminant Function Analysis"; G. David Garson: