[I know XFactor talks about placing just 1 ad unit on your site (above the fold), and an optional 1 link unit on the sidebar, but I have been trying out 3 ad units and 3 link units for the past month. I have noticed a higher CTR while maintaining almost the same amount of cost per click. Based on my own findings, I DON’T believe that placing more AdSense ads on the site would drive the cost per click down. Also, I use #000080 (navy blue) instead of the standard #0000FF blue, and #cccccc for both the description and the URL.]
You have the right idea about testing. What works best on one site does not necessarily transfer to the next site. You need to test each topic to find the optimal placements. Having said that, your testing method may not give you accurate information.
Using anecdotal evidence will lead to many incorrect assumptions. The increase you experienced may have been double if you hadn’t changed the layout. There are so many other external factors that will cause your EPC and CTR to spike or dip for a period of time.
The best approach is to run A/B split tests that are sustained over a period of time. Testing both layouts during the same time period will tend to give you much more reliable data.