Exercise #9
October 10, 2013
Due Tue, Oct 15
Build 2 multiple regression models
Model 1. Analyzing the Database Problem #4 on page 549. You will need the Excel data that we’ve used previously for database exercises: black_7e_excel_database
Model 2. Build a multiple regression model with your set of Pueblo Property Data. See instructions below for Model 2.
- Use the simple “bivariate” regression model developed in Exercise #8.
- Use your previously completed correlation analysis (see Exercise #7) to identify other key variables that may improve the regression model.
- Use the Data Analysis ToolPak in Excel to build several multiple regression models (i.e., models with 2 or more independent variables).
- Save the output from each model run.
- Choose what you consider to be the best model and write a short paragraph describing why you chose it.
Turn in the following (printed) in class on Tuesday:
- Description of Model 1 along with brief answers to questions posed in problem #4 on p. 549
- A list of all model runs attempted for Model 2, include (a) the variables used, (b) the model equations and (c) the resulting r-squared values.
- List the model you’ve selected and a brief written description as to why you think it’s the best.
1 Comment
Here’s a question I received from a student via email. I’m posting here to help others who might be stuck:
“I am trying to complete problem 4 on page 549, and am having trouble knowing how to use the regression analysis.
I am entering Annual Household Income as the y variable and Non Mortgage household debt as the y variable. However when the model generates, I get a value bigger than 1. What am I doing wrong?”
Here’s my answer:
Your dependent variable is y. Dependent means that it needs/depends on x in order to arrive at a corresponding estimated value.
Your independent variable is x. Independent means that we’re not trying to predict/estimate this value. We only want to use it to predict/estimate something else, that something being the dependent variable. In simple bivariate regression there is only one independent variable. In multiple regression there are 2 or more independent variables.
So for problem 4:
Annual Food Expenditure is the dependent variable and should be selected as the y variable.
Annual Household Income and Non-Mortgage Debt are the independent variables that you will use to estimate Annual Food Expenditure and should be selected as the x variable.
Try these selections and let me know how it goes. Hope that helps!