Cluster analysis vs multicolinearity

Hi you all, I hope you can help me with this following problem:

1. I've runned an factor analysis on my data for fifteen "need" items for sugar and sweets, with some knowledge from the company where I work for and background theorie, these items resulted in four factors.

2. Between these four factors, there appears to be still some collinearity. This can be logic because of the fact that people does not have just one single reason to consume sugar and sweets. But know I have to do a cluster analysis with these factors and multicollinearity is one of the assumptions I have to respect.

3. Problem is, how can I reduce the multicollinearity in my data? I found the following solutions:

- Using the Mahalanobis distance measure (not an option for me, SPSS does not have it)
- Cluster variables have to be reduced (not an option because these four are essential to get the right conclusions)
- Reducing the variables to equal numbers in each set of corelated measures (don't know what they mean by this.. source: Hair, Black, Babin, Anderson, Multivariate data analysis (2010)


Thanks in advance for the person who can advise me what do with this..!
Last edited: