Multiple Linear Regression Model

Solution Preview

Please see the attached file.

Babies R'Us Case Study

You have been hired as a strategic consultant by the hugely successful retailer "Babies R'Us". The company sells many products, although one product in particular, a highly innovative car seat, is being test marketed in various domestic and international markets. "Babies R'Us" has hired you to help them better understand the test market data they have compiled from 400 retailers worldwide (found in the excel file called "CarSeats.xls"). The variables in the data set include:

Unit Sales = the number of units sold (in thousands),
Competitor's Price = the price for a similar product being sold by a competitor (in dollars),
Income Level = average household income in the region (in thousands of dollars),
Advertising = amount spent on advertising the product (in thousands of dollars),
Price = the price being charged by your company (in dollars),
Population = number of people (in thousands) living in the region,
Average Age = average age of the population in the region (in years),
Average Education = average educational level in region,
Shelving location = quality of the shelf location for your product (good, medium, or bad), Urban or Rural = description of the region as urban or rural, and finally,
US = a categorical variable indicating whether the sales region is in the US or an international market.

I. Run a multiple regression model with Unit Sales as the dependent variable against all of the available predictor variables, and use that model to answer the questions below. NOTE: DON'T TRY TO ADD OR DEPETE ANY OTHER VARIABLES JUST YET, SIMPLY USE ALL VARIABLES AVAILABLE IN THE ORIGINAL DATA SET:

a. Which independent variables appear to be important predictors of sales? Why?

I'm doing all of this in SPSS (just because it's the software I'm most familiar with). If you don't use SPSS, compare the results you get in Excel, SAS, Systat, or whatever - they should be the same). I changed the categorical variables so that the categories were represented by numbers: Bad = 0, Medium = 1, Good = 2, etc.

Here is the output when I run a multiple regression model:

I put the variables that seems to be a good predictor of sales in bold (these are the variables for which the coefficient - the "B" - is statistically significantly different than 0).

b. Which are not? Any surprises? Explain!

The ones that don't seem to be good predictors of sales are: Population, Years of Education, Urban, and US.

I don't know if any of these are surprising - I don't know anything about car seat ...