Question

The assumption of homoscedasticity requires the residuals (differences between observed and estimated values) to be relatively similar (homogeneous) across different values of the predictor variables. (T/F)

The assumption of normality relates to the distributions of the independent variables; they must be normally distributed. (T/F)

If the distribution of residuals (actual value minus estimated value) is negatively skewed with a mean of 5 and a standard deviation of 1, this indicates that (a) the regression line is estimated below the majority of the data points and (b) there are likely outliers with extremely low values and high leverage on the fit line. (T/F)

As long as the absolute correlation between two independent variables does not exceed .8, multicollinearity is not a concern. (T/F)

Which of the following statistics can be used to evaluate how well a model fits data (select all that apply)?

R-Squared

Adjusted R-Squared

Standardized Beta

Mean Squared Error (MSE)

All of the above

Answer 1

1. The assumption of homoscedasticity requires the residuals to be similar across different values of the predictor variable. Hence, the statement is true.

2. The normality assumption is for the dependent variable, which assumes the dependent variable will be normally distributed. Hence, the given statement is false.

3. If the distribution of residuals (actual value minus estimated value) is negatively skewed with a mean of 5 and a standard deviation of 1, this indicates that (a) the regression line is estimated below the majority of the data points and (b) there are likely outliers with extremely low values and high leverage on the fit line. Hence, the given statement is True.

4. Multicollinearity depends on the VIF values. Hence, the statement is false.

5. R-squared, Adjusted R-squared and MSE can be used to evaluate the model fit.