Re: How to determine if a number is statistically meaningful

From: davegb (davegb_at_safebrowse.com)
Date: 03/24/05 

Date: 24 Mar 2005 06:33:52 -0800

Paige wrote:

I came to the exact opposite conclusion reading the replies. What
you are trying to do *FIRST* requires subject matter expertise, not
statistics. If that's beyond the limits of what you can do, then the
problem is unapproachable.

I never thought to state earlier that I am here as an Excel expert, not
a statistician or even a SME. I was asked to make some spreadsheets
more user friendly, and have done so. One user asked if i could create
a single number to rank the counties' performance, but eliminate the
ones like the example I gave earlier, 4 clients with 1 occurrence.
Being naive, I thought this would be simple. After playing with it on
my own for a while, I realized I couldn't solve it, and came here
thinking that someone would give me a formula or procedure to 1)remove
"statistically insignificant" counties and 2) create a single column of
numbers in the spread*** that would rank the counties performance. We
can all agree now that I was naive.
Having spent quite a bit of time, both in this NG, in 2 separate
threads, I came to the conclusion that I am in over my head, beyond my
pervue. Given time, I'm sure I could understand what I'm being told
here. But I'm supposed to be spending my time making spreadsheets for
others to analyze the data, not on this single problem. And, as I said
above, the solution, if I could come up with one, appears to complex to
be understood by my end users anyway.

Another SERIOUS problem, and the way you state things, it sounds
like you still feel the problem is entirely statistical. It is not.
In any research I have been involved in, defining the problem is the
first step, and normally once those definitions are in place,
everyone comprehends them. Defining the problem clearly *CAN* be
done without resorting to statistical analyses or statistical
methods of any kind.

You're probably right, I have done a poor job of explaining the problem
here. I just don't know what else I could explain to help. Someone
mentioned in one of the threads that if the state wants this done, they
should hire a professional to do it, not some amateur. I agree
completely. A person who understands the problem better as well as some
understanding of the potential choices for a solution would be far
better quailified than myself. I opened a can of worms and lost my
appetite!

Thanks again.