Questions about Correlation, Regression and Confidence Intervals.
(written at Notre Dame to cover material in Chapter 9 of SAS System for Elementary Statistical Analysis)
Click on the answer you think is most nearly correct.
1. Least Squares Regression:
Uses the minimum horizontal distance from the point to the line. Uses the minimum vertical distance from the point to the line. Uses the minimum perpendicular distance from the point to the line. Uses the minimum absolute distance from the point to the line. Minimizes the sum of the deviations from the point to the line.
2. For least squares all variables must be:
continuous or ordinal. nominal or ordinal. nominal or interval. ratio or ordinal. ratio or interval.
3. The term "independent variable" means that:
The variable is not correlated with the error term. The variable is independent of other such variables in the sample. Each explanatory variable is independent of each of the other explanatory variables. The variable is independent of the y variable. The explanatory variable has zero correlation with any other variable in the population.
4. Which of the following statements is not true:
A confidence interval for the mean would be narrower that the confidence interval for a single predicted value. To predict either a single future value or to predict the mean response for a given value of the independent variable, use the same value. To get the confidence limit for the mean use CLM instead of CLI in the PRINT statement. Individual data points can never fall outside of the confidence interval for the mean. Prediction limits take into account both the error in fitting the regression line and the variation between values at a given x while confidence limits for the mean only need to take into account the error in fitting the regression line since the mean doesn't vary.
5. R-square is not:
The ratio of the regression sum of squares divided by the error sum of squares. The coefficient of determination. The multiple correlation coefficient squared. The fraction of the total variation due to the variables in the model. The ratio of the regression sum of squares divided by the total sum of squares.