normalization, normalization, normalization!

Basic Microarray Analysis:

1. Experiment (Raw Data)
2. Image Analysis (Pick Spots)
3. Pre-normalization analysis
   1. Background subtraction
   2. Spatial intensity
   3. Chip to Chip variance
4. Normalization and Filtering
   1. Scale normalization
   2. Intensity dependent location normalization
5. Post-normalization analysis
1. Clustering
2. Hypothesis testing
3. Classification
4. More complicated statistical analysis

I found this list when I was searching the interquartile range (IQR) normalization. Actually Microarray folks might have been done the whole bunch of analysis decades ago. But the question for me is still there: how to compare two samples? 

Let's say, I have two vectors: X=[x1, x2, x3, ..., xn] and Y=[y1, y2, y3, ... yn]. Before I could say X is correlated with Y simply because cor(X, Y) is large, I should make sure that each Xi and Yi are sampled from the same distribution. Right? That's the point. If they are not, I should do normalization ahead to assure that. 

Maybe I should go thru the above list (esp. the normalization part) in some weekend.