Question

17-E2. A survey of 200 randomly chosen consumers was presented with this propositionluxury cars are not worth the money. They rated their degree of agreement with this proposition on a 5-point scale, 5 meaning strongly agree and 2 meaning strongly disagree. There are two main explanatory variables: the respondent’s age and the number of months since the respondent bought a car. Below is some output from two regression models.

Results of multiple regression for expend

Summary measures

Adj R-Square	0.011
Model Error	1.395

Regression coefficients

	Coefficient	Std Err	t-value	P-value	Lower	Upper
Constant	2.220	0.4190	5.30	0	1.393	3.046
Age	0.0154	0.0086	1.79	0.074	-0.0015	0.0323

Results of multiple regression for Response

Summary measures

Adj R-Square	0.008
Model Error	1.397

Regression coefficients

Constant	2.687	0.192	14.00	0	2.308	3.065
months	0.0110	0.0069	1.60	0.111	-0.0025	0.0245

1. Based on these two simple regression models, which is the better predictor of response, and how good is it?
2. Predict the mean response of those customers who last purchased a car 12 months ago.
3. Using multiple regression, describe the effects of age and months since last purchase on consumer response to the proposition. The data are here.
4. Give a predicted response for a 50-year old who last bought a car 3 years ago. How accurate is the prediction?

Answer 1