Variance and ProbF logworth evaluate split worth for which type of variables?

The correct answer highlights that variance and ProbF logworth are primarily applicable to interval variables. Interval variables are characterized by numerical values where the differences between values are meaningful and can be used in mathematical calculations. This includes variables such as temperature or test scores, where both the distance and the ordering matter, allowing for variance to be calculated effectively.

In this context, variance helps assess how much the data points within a group differ from the mean, giving insights into the distribution of the data. The ProbF logworth is linked to statistical significance in the context of testing hypotheses or determining the worth of a split in decision trees, and it utilizes variance to help indicate how informative a particular split is.

Other types of variables, such as categorical, ordinal, and dichotomous, do not fit this evaluation method as they have different characteristics. Categorical variables represent distinct groups without inherent numerical values, ordinal variables suggest a ranking but without consistent intervals, and dichotomous variables indicate a binary outcome. These variable types require different approaches for analysis that are not centered on the calculations of variance and ProbF logworth.