Chapter 14, Correspondence Analysis Video Solutions, Statistics for Marketing and Consumer Research

Problem 1

Open the Trust data-set
a. Run a simple correspondence analysis between q54 (job status) and q61 (financial situation of the household)
b. Show the bi-plot for the symmetric, row principal and column principal models - which one provides better result?
c. Which dimension is most important in representing points for unemployed and fully employed individuals?
d. Is the total inertia acceptable? Which category is least explained by the model?

Check back soon!

04:40

Problem 2

Open the EFS data-set
a. Using simplecorrespondence analysis plot tenure type (a121) and household composition (a062)
b. Show the bi-plot
c. If the Euclidean distance is used instead of the Chi-square one, do results change?
d. Check the total inertia - is its value acceptable?
e. Which category is best described in the first dimension? Which one is worst described?
f. Try and look for a third variable and draw the tri-dimensional plot

Amany Waheeb

Numerade Educator

Problem 3

Open the EFS data-set
a. Your objective is to reduce into two components the following categorical variables - a062 (household composition) a121 (tenure type) gor (region) a093 (economic position) a094 (NS-SEC8 class)
b. Use multiple correspondence analysis (DATA REDUCTION/OPTIMAL SCALING) to extract two components and save the scores
c. Plot tenure type with region and economic position
d. Use gender (a071) as supplemental variables. How do results change?
e. Draw a table with the average scores by gender

Check back soon!