Search All of the Math Forum:
Views expressed in these public forums are not endorsed by
Drexel University or The Math Forum.



Stratified versus Block (fwd)
Posted:
Nov 25, 1996 11:02 AM


Bob 
Thanks for remembering one of my favorite expressions. My 1995 summer workshop participants gave me an endofsession gift of a T shirt that says:
"Can't Conrol It  Measure It"
 Joe
 Forwarded message  Date: Thu, 14 Nov 1996 21:39:06 0500 (EST) From: Bob Hayden <hayden@oz.plymouth.edu> To: "AP Stats. list" <apstatl@etc.bc.ca> Subject: Stratified versus Block
 Forwarded message from AlCoons@aol.com 
We are trying to distinguish between Stratified Random Sampling and Block Design. BPS makes the point:
"Blocks and strata both group similar units. We use two different names only becuase the idea developed separately for sampling and experiments." pg 214
So what is the party line? Are they the same? Is there a difference?
Thanks in advance,
Al
 End of forwarded message from AlCoons@aol.com 
I'd say it is the same concept applied to different situations. Here's another example. When I teach linear models I present analysis of covariance in the following way.
Sometimes we want to study the effect of one or more variables x1, x2, ... on another variable, y. We ALREADY know that y also depends on z1, z2, ..., and if we do not include the z's in the model, we will get a very poor model. So we include the z's to get a good model, and then test hypotheses about the x's.
**************************************************************************** Or as Joe Ward would say, "If you can't control for it, measure it." *****************************************************************************
In the traditional situation, the x's represent cells in an ANOVA and there is one z and it is a measurement variable. So, is what I do REALLY analysis of covariance? I'd say it is the same idea applied to a broader range of situations. I prefer the broader framework for teaching, and see no reason to say much about the narrower idea. However, if you were going to do a literature search, it might be useful to know that virtually all of the literature you turned up would be about the narrower situation.
Similarly, if you are teaching blocking and stratified sampling, I don't think it is important to draw a distinction. However, if you want to do a literature search it would be useful to know that the history of this basic idea within experimental design has been different than its history within survey sampling, so you might need to search both areas, using the locally accepted terminology in each, if you wanted to research the impact of this basic idea in statistics.
But for purposes of AP Stats., I think what David Moore says is exactly what should be said on this topic (indeed, on MOST topics!).
_   Robert W. Hayden   Department of Mathematics /  Plymouth State College MSC#29   Plymouth, New Hampshire 03264 USA  *  Rural Route 1, Box 10 /  Ashland, NH 032179702  ) (603) 9689914 (home) L_____/ hayden@oz.plymouth.edu fax (603) 5352943 (work)



