• Home
  • Textbooks
  • Statistics for Marketing and Consumer Research
  • Correlation and Regression

Statistics for Marketing and Consumer Research

Mario Mazzocchi

Chapter 8

Correlation and Regression - all with Video Answers

Educators

Chapter Questions

01:34

Problem 1

Open the EFS data-set
a. Explore the bivariate correlation matrix for the following variables
i. Expenditure in petrol (c72211)
ii. Expenditure in railway and tube fares (exluding season; c73112)
iii. Expenditure on taxis (c73212)
iv. Anonymized household income (incanon)
b. Compute the Spearman's Rho and the Kendall's Tau b - are they very different from Pearson's bivariate correlations?
c. Income is probably influencing the relationship between expenditure in petrol and expenditure on taxis - use partial correlations to account for such an effect
d. Now look at the partial correlation between taxis and petrol after accounting for both income and railway/tube fares
e. What do the above corrrelation values mean? Which ones are significant?

Kaylee Mcclellan
Kaylee Mcclellan
Numerade Educator

Problem 2

On the EFS data-set
a. Run a bivariate regression where expenditure on petrol (c72211) is the dependent variable and income (incanon) the explanatory variable
b. Compare the results with the bivariate correlations of previous exercise
c. Generate the residuals from the regression (hint - click on SAVE in the regression dialog box and ask for unstandardized residuals - a new variable is created)
d. Plot the residuals into a histogram. Do they look normally distributed with mean zero? Perform a test for normality among those suggested in chapter 6 .

Check back soon!
04:35

Problem 3

Open the Trust data-set
a. Run a multiple regression where expenditure on chicken (q5) is the dependent variable and the explanatory variables are:
i. Price (price)
ii. General attitude toward chicken ( $q 9)$
iii. Perceived risk from chicken consumption (q27d)
b. What is the goodness-of-fit?
c. Estimate the model again and add the income variable (income) and the number of household components (q56). What is the goodness-of-fit now? Which goodness-of-fit indicator allows comparison between the two models?
d. Estimate the regression equation again with the stepwise method. Which variables are kept in the model?
e. Interpret the coefficients,relating unitary changes in the explanatory variables to changes in the dependent variables.

Sophie Knight
Sophie Knight
Numerade Educator