Say I have some data with observations of individuals with different treatment. I can test whether the treatments affect the observations with an ANOVA test.
But now I want to find out whether the variance in the observations within treatments is affected by some other variable. If the treatment were a continuous variable, I could do (in R)
fit
The examples used in the ncvTest function in the car package include categorical variables. I'm not sure if this addresses your specific case.
Modelling iinvolving regression requires that all variables in use. are continous variables. Categorical variables will be treated as dummy variables. An alternative approach is to do principal component analysis on the categorical variables, and then substitute their factor scorethe original categorical variables. Factor scores from principal component analysis are usually interval data or continuous variables.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Statistical Software
- R Statistical Package