[NatureNS] Estimating (guessing) numbers of birds

Date: Mon, 08 Jul 2013 12:12:53 -0300
From: iamclar@dal.ca
To: naturens@chebucto.ns.ca
User-Agent: Internet Messaging Program (IMP) H3 (4.3.4)
Precedence: bulk
Return-Path: <naturens-mml-owner@chebucto.ns.ca>
Original-Recipient: rfc822;"| (cd /csuite/info/Environment/FNSN/MList; /csuite/lib/arch2html)"

next message in archive
next message in thread
previous message in archive
Index of Subjects


All:

Nancy's dilemma and suggestions on estimating numbers for eBird are  
important. I have been a frequent user of eBird, atlas, CBC, BBS,  
etc., over the years in estimating trends and patterns,. I believe the  
dictum of a statistician colleague that "a large number of bad numbers  
(as long as not biassed) are as good as a small number of good  
numbers" (a recasting of the "law of large numbers"). Even the casual  
reporting of numbers like "lots" or "common" on naturens are less  
useful for section editors of "NS Birds" than are estimates.

Nancy suggests that a range of categories could be available on eBird  
2-10, 11-50, etc., but the "users" of these numbers would still have  
to turn these into a single number for trend analyses, etc. I believe  
that an educated guess is better; Why not just take the rough number  
between 11 and 50, e,g., that you think might be better - 20 or 40?

Another possible guide might be that used during the first (1980s) NS  
Atlassers of entering log categories 1-10, 10-100, etc. (natural logs  
even better). Then (as was done) the user (or indeed the observer) can  
use the geometric mean (about 3, about 30, etc.).

Cheers, Ian

Ian McLaren

next message in archive
next message in thread
previous message in archive
Index of Subjects