Tex, I have no idea. As I said earlier, I'm not a statistician. It doesn't sound like you are either. Perhaps you should write to the authors of the original study and ask for their response? You could also try asking the folks at STATS.
You read can the actual report here - http://www.thelancet.com/webfiles/im...3606694919.pdf
BTW - Your Katrina story proves nothing. Just because one study using a particular methodology is wrong doesn't mean the method is.
As an example: Let's say a colleague and I use radio-carbon dating to determine the age of two different archaeological artifacts from two different archaeological sites. It is later proven that I made a few mistakes taking my sample for dating and it threw off my dating by several thousand years. This doesn't make my more careful colleague’s results (and every other Radio Carbon date) null and void simply because he was using the same dating method. It just shows I was careless. You see where I'm going?
Now I'm not qualified to debate the intricacies of statistical sampling but from what I've read by experts in the field it appears cluster sampling is an accepted technique. This particular study could be wrong but I've yet to see any credible scientific critique demonstrate that.
EDIT -I have never said that nor implied it. I am quite familiar with what statistics are.You can't use statistics as facts, as iron seems to be implying here
Bookmarks