Chapter 16: Regression Analysis: Model Building

Model building is solely concerned with identifying which independent variables should be included in the model.
1. True
2. False
A residual plot may be helpful in determining if a curvilinear relationship is appropriate.
1. True
2. False
The t test for determining the significance of adding only one additional independent variable to an existing model is equivalent to the partial F test.
1. True
2. False
With the stepwise regression procedure, a variable is added at each step, but variables are never deleted.
1. True
2. False
One may use a logarithmic transformation involving the dependent variable
1. an exponential model is appropriate.
2. the residual plot suggests that the variance of the error term is not constant.
3. Both of the above are correct.
4. None of the above
At each step of the stepwise regression procedure, the first consideration is to see whether any variable can be added.
1. True
2. False
When autocorrelation is present, at least one of the assumptions of the regression model is violated.
1. True
2. False
The Durbin-Watson test is generally inconclusive for small sample sizes.
1. True
2. False
The model y = b_o + b₁x₁ + b₂x₁² + eis called a
1. simple nonlinear model with one predictor variable
2. simple first-order model with one predictor variable
3. second-order model with one predictor variable
4. second-order model with two predictor variables
Interaction terms are added to a regression model in order to
1. study the joint effect of quantitative and qualitative variables
2. account for curvilinear effects in the data
3. transform a nonlinear model into an equivalent linear model
4. account for the joint effect of two independent variables
The F test for determining whether to add two variables to an existing model is based on a determination of
1. the increase in SSE resulting from adding the two variables
2. the decrease in SSR resulting from adding the two variables
3. the change in SST resulting from adding the two variables
4. the decrease in SSE resulting from adding the two variables
5. none of the above
The purpose of the Durbin-Watson Test is to test
1. for a significant difference between a full model and a reduced model.
2. for autocorrelation.
3. the significance of an individual independent variable.
4. the significance of the model.
5. None of the above
The purpose of the partial F test is to test
1. for a significant difference between a full model and a reduced model.
2. for autocorrelation.
3. the significance of an individual independent variable.
4. the significance of the model.
5. e. none of the above
The partial F test is used to compare a full model to a reduced model. A way to choose an appropriate reduced model is to use
1. a. backward elimination.
2. forward selection.
3. stepwise regression.
4. All of the above are correct
5. None of the above
The multiple regression approach to analysis of variance and experimental design uses
1. autocorrelation.
2. dummy variables.
3. interaction.
4. logarithmic transformations.
5. None of the above

Chapter 16: Regression Analysis: Model Building

Quiz