Chapter 12, Cluster Analysis Video Solutions, Statistics for Marketing and Consumer Research

Problem 1

Open the Trust data-set
a. Segment the sample into four clusters according to the attitude variables ( $\mathrm{q} 12 \mathrm{a}$ to $\mathrm{q} 12 \mathrm{k}$ ), using the $k$-means algorithm
b. Save the final seeds in a file
c. Replicate the analysis using the final seeds as initial seeds
d. Perform an ANOVA analysis to check whether the clusters are significantly different with respect to income (q60) and age (q51)
e. Extract three principal components from the attitude variables, save the component scores and run the $k$-means algorithm on the component score
f. Do the cluster averages differ from the previous ones with respect to attitudes, income and age?

Check back soon!

Problem 2

Open the EFS data-set
a. Segment the UK population with respect to household expenditure in various goods (use expenditure variables from p600 to p612)
b. What is the optimal number of cluster according to the Ward algorithm?
c. Run a $k$-means cluster analysis using the number of clusters chosen at point $b$
d. Save the cluster-membership from the $k$-means clustering
e. Run cluster analysis again, using a hierarchical algorithm and choosing the same number of clusters as before
f. Save the cluster-membership from the hierarchical algorithm
g. Show the averages of p600 to p612 for both clustering methods. How different are they?

Check back soon!

01:39

Problem 3

Open the Trust data-set
a. Use SPSS two-step clustering to identify a number of clusters not larger than five on the basis of the risk variables ( $\mathrm{q} 27 \mathrm{a}$ to $\mathrm{q} 27 \mathrm{~g}$ )
b. Compute the average of the risk aversion variable $q 28$ within each cluster
c. Do the cluster means of $\mathrm{q} 28$ show significant differences across clusters?
d. Run the analysis again choosing the number of clusters with the Ward algorithm and run the actual classification with $k$-means
e. Compare the results

Dominador Tan

Numerade Educator