Just a comment about data analysis. Over 15 years ago I had reason to analyze a 120 area data set, with maybe 20 or less variables. I called a professional data guy ($350 or so each time he helped me, as I recall) and he first said to spend $400 or so on a stat package, as I recall, to get a data analysis package. For a grand or less I had access to consultation on the use of the pack and interpretation of the print outs. Did it all for you. You tossed in the data set and chose which stats you wanted it to do. Bang. Done. The point I am making is that data analysis is done by stat packages, quickly and relatively cheaply. Of course you need a pro to pick out the package, help with the choice of which tests to use and how to interpret results. But it is not rocket science and it is relatively cheap. All the fuss about secondary data analysis after a failed test is ingenuous, in my view. These are small sample sizes with not that many variables. The final prices should be thousands or tens of thousands. Not a crushing expenditure. After all, we are not dealing with patients, just numbers.