# Simple Questions

#### jamesmartinn

##### Member
...and yet i'm stuck.

1) How would you represent the components of ANOVA of regression in a graphic format,when considering a bivariate regression? How would you do describe these components for two predictors?

Basically in the context of the ANOVA table wrt to Regression:
SSTotal = total variation of Y scores around the mean of Y
SSRegression = represents the sum of squared distance between the predicted scores of Y by the regression line Yhat and the grand mean of Y
SSerror = represents the total variation of points and the corresponding points on the regression line.

I'm having some trouble translating this into the multiple regression case with 2 predictors

2)When is it appropriate to perform a transformation? What is the justification for performing a transformation? (In other words, why is it okay to perform a transformationon data?). What is the connection between the transformed sample, the sampling distribution and the population distribution?

A transformation is appropriate if your original scores do not conform to assumptions of the test you plan to use (i.e, normality, homogeneity of variances).