Subject: | |
From: | |
Reply To: | |
Date: | Tue, 26 Aug 2003 12:05:28 -0400 |
Content-Type: | text/plain |
Parts/Attachments: |
|
|
It is a statistical matter - and looking at mean dates alone does not tell
us much. If artifacts are going to be quantified and tied to dates or
regression formulas, Then the whole range or distribution of data needs to
be accounted for in an analysis, even the outliers. The peaks, spaces, and
outliers can be more significant than the mean, mode, trimmed mean, etc…
Grant
-----Original Message-----
From: HISTORICAL ARCHAEOLOGY [mailto:[log in to unmask]]On Behalf Of chris
rohe
Sent: Tuesday, August 26, 2003 10:29 AM
To: [log in to unmask]
Subject: Re: TPQs etc
This is mainly a statistical matter and probably an often misuse of
statistics. A mean is meant to give an idea about a data set not really
predict anything- can be if the data are normally distributed. If you have
a group of data with breaks in it then the mean nay well fall in one of
those breaks or be pulled left or right due to outliers. Shouldn't you just
use your latest dated artifact within a certain context anyhow? Maybe using
the mode may be more informative.
|
|
|