Date: Tue, 16 Mar 2004 19:31:01 -0500 Reply-To: Maryland Birds & Birding Sender: Maryland Birds & Birding From: Bill Ellis Subject: Re: Scoter numbers In-Reply-To: <003f01c40baa$00d743f0$6401a8c0@HERKY> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit More seriously than semi-seriously: "Significant figures" covers the set of informal rules for keeping measured quantities sensible. As gut instinct would tell you, adding together "about 41,000" and 22 gives 41,000. "About 41,000" implies that the number counted had only 2 significant figures (4 and 1), or possibly 3 significant figures (4, 1, and the first 0). When adding or subtracting, the answer cannot have a significant figure in a place (hundred, tens, units, tenths, hundredths, etc.) lower than all the numbers had. (Expressed another way, the precision of the result cannot exceed the precision of the least precise contributing number.) If the 41,000 scoters came from counting them one at a time, the number would have 5 significant figures, including one in the units position (a 0), and the sum of 41,000 and 22 would be exactly 41,022, which also has a significant figure in the units position. Because the zeros in a number may or may not be significant figures, scientists use "scientific notation" in reporting measurements so that there is no doubt about how many significant figures they are claiming: 41,000 to the nearest thousand would be 4.1 x 10exp4 (10 raised to the 4th power) 41,000 to the nearest hundred would be 4.10 x 10exp4 Exactly 41,022 would be 4.1022 x 10exp4 When bird counts are reported at a tally, the sum of all subregions counts should only have a digit in a place that was significant for every subregion counted. If a region reports 100, the tally recorder needs to ask how many places were significant - if it was a single flock, the answer is probably 2. In physics, it is not unusual to have measurements in careful, elaborate experiments reported with eight significant figures. In carefully measured environmental pollutant reports (analytical chemistry), 3 to 4 significant figures is all that can be justified. Bird counts are similar. George: My experience with one PhD statistician was that he never heard of significant figures. Scientists usually use them faithfully. Bill Ellis Eldersburg Carroll County -----Original Message----- From: George M. Jett Sent: Tuesday, March 16, 2004 5:57 PM Jim I will ask a PhD statistician friend of mine if you can have an exact number which is composed of estimates and actual numbers. Looks goofy to me. ... ----- Original Message ----- From: "JAMES FELLEY" Sent: Tuesday, March 16, 2004 9:11 AM > George Jett wondered: > > >How can you have an exact number of "155,422 scoters" if > >you have estimated the number of Black Scoters as > >"About 41,000"? > ... > Why disrespect the White-winged Scoters by ignoring > them in the total count? > > (semi-seriously) Jim ======================================================================= To leave the MDOsprey list, send e-mail to listserv@home.ease.lsoft.com with the following message in line 1: signoff mdosprey ======================================================================= =========================================================================