Support in interpreting PCA output

#1
Hi all,

I have read quite a bit about PCA on the web and also the post written by Terzi.
I think I have understood the concepts behind it but I need some help in interpreting the results. I have found useful the post from Terzi but I got lost when he shifted from the first to second post.

I am currently using PCA to create a wealth index for some states.
All the variables I am using are dichotomous, dummy and they are about whether a household has or not a certain item (TV, electric fan, etc).

This is the output that I get:



Uploaded with ImageShack.us

Ok, so I have tried to read here and there how to read such an output.
So basically the first three components have the largest eigenvalue, therefore, are the ones that explain the biggest variance.
According to the screeplot, I should consider electricity in the household, cell phone and pressure cooker as the principal component factors.
However, in terms of proportion, they seem to explain very little.
Does it mean that having electricity in the house (elhh) can explain only the 0.26 % of the variance?

I am really confused!


....thanks a lot !!

belfagor
 
#2
Ok so, O have understood that 0.26 means actually 26%. Yet, I was expecting more as this is for electricity in the house.
Should I add more variables given the low % that I get?
Plus, how do I now go from here to scoring factors and then to an index?