Multiple correspondence analysis

Hi, my survey data are about product preferences (respondents choose 3 most favorite ones from 16 products) and product qualities (one quality from 6 options for each chosen product). These variables I call dependent variables.
I want to know how these responses are influenced by independent variables - socio-demoghraphic characteristics of respondents.

First I decided to reduce number of dependent variables.
So, I conducted the multiple correspondence analysis as the data are categorical. But the resulting model has many dimensions each of them explaining around 10% or less of variance. Which outputs help me to appropriately "cluster" the dependent variables?
Should I look at correlations of original variables?
I am using the SPSS,

Thanks
Lone