Subject: ECOLOG-L Digest - 7 May 2003 to 8 May 2003 (#2003-122)
There are 24 messages totalling 1139 lines in this issue.
Topics of the day:
1. Statistics teaching - out on a limb
2. Native grasses & forbes, neNA
3. addition of standard errors
4. ANN: Course Multivariate Analysis of Ecological Data
5. Dickerson, et al (1997)
6. Computational ecology postdoctoral positions at Tennessee
7. citation for Var(X-bar*Y-bar)
8. Statistics teaching (2)
9. adding variances - related question...
10. Towards Best Practices -- An On-line Forum
11. posting of pdf versions of manuscripts on websites/ stats discussion (3
12. US to open public wilderness lands for...
13. Combining standard errors (2)
14. [Ecolog] citation for Var(X-bar*Y-bar) {01}
15. Two unrelated answers
16. Aquatic Ecology Graduate Assistantship, Univ. of IL-Springfield
17. US to open public wilderness lands .. (2)
18. ECOLOG--posting PDF files to web
19. posting of pdf versions of manuscripts on websites
----------------------------------------------------------------------
Date: Wed, 7 May 2003 21:47:48 -0500
From: "Klawinski, Paul" <klawinskip@WILLIAM.JEWELL.EDU>
Subject: Re: Statistics teaching - out on a limb
Bill and others,
If you are referring to standard devations as your measure of error,
wouldn't you add them up. As I understand it, the amount of uncertainty in
a mean computed from a variety of other means is increased by each mean's
associated uncertainty.
Paul
Paul Klawinski, Ph.D.
Assistant Professor
Department of Biology
William Jewell College
500 College Hill
WJC Box 1040
Liberty, MO 64068-1896
816.781.7700 ext 5568
klawinskip@william.jewell.edu
> -----Original Message-----
> From: William Silvert [mailto:wsilvert@MILPAH.COM]
> Sent: Wednesday, May 07, 2003 2:21 PM
> To: ECOLOG-L@LISTSERV.UMD.EDU
> Subject: Re: Statistics teaching
>
>
> David asks a fair question, and he is not the only one who
> has asked me
> about this. The answer should be in any decent stats book (I
> learned it as a
> college freshman in a physics class from "Theory of Error" by
> Yardley Beers,
> a truly fine little book), but if you want to play around with the
> calculation I have a simulation (done in Excel, you should
> forgive me) which
> will be available in a day or so at the URL
> http://silvert.org/error_propagation.zip. But I am curious to
> know how many
> of the subscribers to this list actually have had the chance to learn
> anything about error propagation, i.e., how to add (or multiply) two
> numbers?
>
> Bill Silvert
>
> ----- Original Message -----
> From: "David M Bryant" <dmb@io.harvard.edu>
> To: "William Silvert" <wsilvert@MILPAH.COM>;
> <ECOLOG-L@LISTSERV.UMD.EDU>
> Sent: Wednesday, May 07, 2003 7:16 PM
> Subject: Re: Statistics teaching
>
>
> > I have asked several statisticians this same question "
> i.e. how does one
> > combine the errors of two means" and was told "most
> statistics programs
> > should handle it". Which I took to mean " I don't know but
> I'm not going
> > to admit it to you". The somewhat pejorative tone of your
> statement below
> > implies that it is an easy task. Yet I can recognize the
> problem that
> > adding the mean of apples and oranges to come up with the
> mean number of
> > fruit is not straight forward, not to mention the error terms.
> >
> > So I'm asking you to conform to my personal credo that no
> criticism should
> > be given without a workable solution. Therefore, how does
> one add the
> > means of apples and oranges to come up with the mean fruit AND the
> standard
> > error?
> >
> > David
> >
> >
> > At 12:01 PM 5/7/2003 +0100, William Silvert wrote:
> > >This brought to mind a sore point with me, that many ecologist
are
> familiar
> > >with very sophisticated statistical techniques, but not
> with the basics.
> I
> > >have a lot of horror stories from conferences and papers I
> have reviewed,
> > >but this is an example of what I mean. I recently posed a ques
ion
> similar
> > >to this to a large group of marine ecologists (I changed
> it a bit here to
> > >use more common species):
> > >
> > >You are studying an area in which the only invertebrates
> are snails and
> > >earthworms. The densities are 10+_3 snails/m2 and 20+_4
> earthworms/m2
> (by
> > >+_ I mean the standard plus or minus error). Assuming that
> the species
> are
> > >independent, how many inverts are there per square meter?
> > >
> > >It is a simple problem in propagation of error, which only
> one person
> could
> > >solve. Given that most experiments require combining
> several observations
> > >which have associated errors, it seems to me that learning
> how to add two
> > >numbers together should preceed the use of advanced statistics
> > >
> > >Bill Silvert
> > >
> > >----- Original Message -----
> > >From: "Hamachan Hamazaki" <hamachan_hamazaki@fishgame.state
ak.us>
> > >To: <ECOLOG-L@LISTSERV.UMD.EDU>
> > >Sent: Tuesday, May 06, 2003 8:12 PM
> > >Subject: Re: Alternative to EXCEL stats
> > >
> > >
> > > > More importantly, I hope some one is teaching students
> how to use
> those
> > > > statistics programs and interpretation of its results.
> Recently, I
> taught
> > > > biostatistics to first year biology graduate students.
> Despite the
> fact
> > > > they have taken Stats course in undergraduate, none of t
em had
> experience
> > > > in using a statistics software. Nor did they have
> experience in just
> > > > cleaning up and manipulating datasets to make
> statistical analyses
> > >possible.
> > > >
> > > > Toshihide Hamazaki / Hamachan
> >
> >
> >
>
------------------------------
Date: Thu, 8 May 2003 08:34:56 -0400
From: eann@JUNO.COM
Subject: Native grasses & forbes, neNA
I'm seeking information (e.g., species lists, ranges, hardiness, growing
requirements, herbivory, etc.) re: native grasses and forbes in
northeastern North America. Can someone recommend a personal contact or
reference materials?
I appreciate your help,
Ann
~*~ ~*~ ~*~ ~*~ ~*~ ~*~ ~*~ ~*~ ~*~ ~*~ ~*~ ~*~
E. Ann Poole, M.Sc., Ecologist & Environmental Planner
479 N State St, F-2, Concord, NH 03301
<eann@juno.com> 603.230.9870
~*~ ~*~ ~*~ ~*~ ~*~ ~*~ ~*~ ~*~ ~*~ ~*~ ~*~ ~*~
------------------------------
Date: Thu, 8 May 2003 09:14:39 +0200
From: Martin Koechy <martin.koechy@FREENET.DE>
Subject: Re: addition of standard errors
It's called propagation of errors. Have a look at
http://mathworld.wolfram.com/ErrorPropagation.html
Martin
--
Martin Köchy (Koechy), PhD
Telefon: +49-331-977 1974 Fax: +49-331-977 1948
[ http://home.wtal.de/koechy/ ] [ http://www.bio.uni-potsdam.de/vegnat ]
AG Vegetationsökologie & Naturschutz, Institut für Biochemie & Biologie
(RG Vegetation Ecology & Nature Conservation, Biochemistry & Biology Dept.)
Universität Potsdam * Postfach 60 15 53 * 14415 Potsdam * GERMANY/ALLEMAGNE
------------------------------
Date: Thu, 8 May 2003 09:40:27 +0200
From: Petr Smilauer <petrsm@JCU.CZ>
Subject: ANN: Course Multivariate Analysis of Ecological Data
I would like to announce the next run of our course (Multivariate
Analysis of Ecological Data). It runs in January 2004
and its main focus are the constrained ordination methods (CCA, RDA)
used with the Canoco software, but we spend substantial time with
other topics (other ordination methods, including NMDS, regression
methods like GLM and GAM, TWINSPAN, cluster analysis, etc).
The participants spend large part of this ten days course working with
their own data, presenting their "mini-projects" at the end of course.
Students pay substantially lower fee for our course.
The course complements our book to be published next month at Cambridge
University Press:
http://www.cup.org/titles/catalogue.asp?isbn=0521891086
Course quota is 20 participants and we have already 12 seats booked.
If you are interested, I suggest you first check the course' web pages
at: http://regent.jcu.cz
Best regards
Petr Smilauer
Faculty of Biological Sciences
Ceske Budejovice
Czech Republic
------------------------------
Date: Thu, 8 May 2003 10:29:57 -0400
From: eann@JUNO.COM
Subject: Dickerson, et al (1997)
I'm requesting assistance to find the following book:
Dickerson, J., D. Burgdorf, T. Bush, C. Miller, B. Wark, R. Maher, and B.
Poole. 1997. Vegetating With Native Grasses in Northeastern North
America. USDA/NRCS Plant Materials Program and Ducks Unlimited Canada,
Barre, Ontario. 134pp + appendices
Thanks, again,
Ann
~*~ ~*~ ~*~ ~*~ ~*~ ~*~ ~*~ ~*~ ~*~ ~*~ ~*~ ~*~
E. Ann Poole, M.Sc., Ecologist & Environmental Planner
479 N State St, F-2, Concord, NH 03301
<eann@juno.com> 603.230.9870
~*~ ~*~ ~*~ ~*~ ~*~ ~*~ ~*~ ~*~ ~*~ ~*~ ~*~ ~*~
------------------------------
Date: Thu, 8 May 2003 10:26:42 -0400
From: Louis Gross <gross@TIEM.UTK.EDU>
Subject: Computational ecology postdoctoral positions at Tennessee
Postdoctoral Positions at the University of Tennessee
May 2003
Two one-year positions are available immediately, based in The Institute
for Environmental Modeling and the Departments of Ecology and
Evolutionary Biology and Mathematics in Knoxville, Tennessee.
1. Mathematical biologist/applied mathematician to collaborate on new
approaches in spatial optimization and control of ecological and
epidemiological models under the direction of Dr. Suzanne Lenhart and
Dr. Louis Gross.
2. Computational ecologist to collaborate on ecological modeling projects
involving multimodeling of Everglades restoration and linkages to parallel
computation and grid-computing under the direction of Dr. Louis Gross.
Interested individuals should contact Dr. Gross (gross@tiem.utk.edu)
for further information. See http://www.tiem.utk.edu/ for information
about The Institute.
------------------------------
Date: Thu, 8 May 2003 09:27:53 -0500
From: "Larson, Michael Andrew" <LarsonM@MISSOURI.EDU>
Subject: citation for Var(X-bar*Y-bar)
Not long ago I had difficulty finding the proper way to calculate the varian
e of a product; I was calculating the fecundity terms (fertility*survival) f
r a population projection (i.e., Leslie) matrix. As some have mentioned ear
ier on this thread, many basic stats texts do not provide such equations for
propagation of error. The lone exception seems to be that most of the texts
I have provide guidance for the SE of a difference between two means.
For those who may be interested,
Var(X*Y) = [X^2 * Var(Y)] + [Y^2 * Var(X)] - [Var(X) * Var(Y)]
where Var=variance, X=mean of sample 1, Y=mean of sample 2, and X^2=X-square
.
Please check the original publication in case I copied something incorrectly
Goodman 1960 J. Amer. Stat. Assoc. 55:708-713.
Mike Larson
Fisheries & Wildlife
University of Missouri
------------------------------
Date: Thu, 8 May 2003 11:42:24 -0700
From: brian@UIDAHO.EDU
Subject: Statistics teaching
The fourth edition of "Introduction to the practice of statistics", by Moore
& McCabe states, on p. 330:
Rule 2. If X and Y are independent random variables, then
Var(X+Y) = Var(X) + Var(Y)
More general error propagation problems, such as for products, ratios, & sit
ations with dependent random variables, require concepts from elementary mat
ematical statistics. I have long advocated that ecologists who are heavy st
tistics users should take a one year math-stat course sequence in their trai
ing. Such a sequence is normally offered at the senior-year undergraduate l
vel at most universities, titled something like "Intro to probability" and "
ntro to mathematical statistics"; an example textbook for the sequence is:
Mathematical statistics and data analysis, second edition" by J. A. Rice.
With this course sequence, the ecologist would have to take LESS statistics
ourses overall (yes, really!) in order to become MORE confident and skillful
in the use of statistics.
Brian Dennis
Professor of Wildlife and Statistics
University of Idaho
----- Original Message -----
From: William Silvert <wsilvert@MILPAH.COM>
Date: Wednesday, May 7, 2003 12:21 pm
Subject: Re: Statistics teaching
> David asks a fair question, and he is not the only one who has
> asked me
> about this. The answer should be in any decent stats book (I
> learned it as a
> college freshman in a physics class from "Theory of Error" by
> Yardley Beers,
> a truly fine little book), but if you want to play around with the
> calculation I have a simulation (done in Excel, you should forgive
> me) which
> will be available in a day or so at the URL
> http://silvert.org/error_propagation.zip. But I am curious to know
> how many
> of the subscribers to this list actually have had the chance to learn
> anything about error propagation, i.e., how to add (or multiply) two
> numbers?
>
> Bill Silvert
>
> ----- Original Message -----
> From: "David M Bryant" <dmb@io.harvard.edu>
> To: "William Silvert" <wsilvert@MILPAH.COM>; <ECOLOG-
> L@LISTSERV.UMD.EDU>Sent: Wednesday, May 07, 2003 7:16 PM
> Subject: Re: Statistics teaching
>
>
> > I have asked several statisticians this same question " i.e. how
> does one
> > combine the errors of two means" and was told "most statistics
> programs> should handle it". Which I took to mean " I don't know
> but I'm not going
> > to admit it to you". The somewhat pejorative tone of your
> statement below
> > implies that it is an easy task. Yet I can recognize the
> problem that
> > adding the mean of apples and oranges to come up with the mean
> number of
> > fruit is not straight forward, not to mention the error terms.
> >
> > So I'm asking you to conform to my personal credo that no
> criticism should
> > be given without a workable solution. Therefore, how does one
> add the
> > means of apples and oranges to come up with the mean fruit AND the
> standard
> > error?
> >
> > David
> >
> >
> > At 12:01 PM 5/7/2003 +0100, William Silvert wrote:
> > >This brought to mind a sore point with me, that many ecologist
are
> familiar
> > >with very sophisticated statistical techniques, but not with
> the basics.
> I
> > >have a lot of horror stories from conferences and papers I hav
> reviewed,> >but this is an example of what I mean. I recently
> posed a question
> similar
> > >to this to a large group of marine ecologists (I changed it a
> bit here to
> > >use more common species):
> > >
> > >You are studying an area in which the only invertebrates are
> snails and
> > >earthworms. The densities are 10+_3 snails/m2 and 20+_4
> earthworms/m2(by
> > >+_ I mean the standard plus or minus error). Assuming that the
> speciesare
> > >independent, how many inverts are there per square meter?
> > >
> > >It is a simple problem in propagation of error, which only one
> personcould
> > >solve. Given that most experiments require combining several
> observations> >which have associated errors, it seems to me that
> learning how to add two
> > >numbers together should preceed the use of advanced statistics
> > >
> > >Bill Silvert
> > >
> > >----- Original Message -----
> > >From: "Hamachan Hamazaki" <hamachan_hamazaki@fishgame.state
ak.us>
> > >To: <ECOLOG-L@LISTSERV.UMD.EDU>
> > >Sent: Tuesday, May 06, 2003 8:12 PM
> > >Subject: Re: Alternative to EXCEL stats
> > >
> > >
> > > > More importantly, I hope some one is teaching students h
w
> to use
> those
> > > > statistics programs and interpretation of its results.
> Recently, I
> taught
> > > > biostatistics to first year biology graduate students.
> Despite the
> fact
> > > > they have taken Stats course in undergraduate, none of t
em had
> experience
> > > > in using a statistics software. Nor did they have experi
nce
> in just
> > > > cleaning up and manipulating datasets to make statistica
> analyses> >possible.
> > > >
> > > > Toshihide Hamazaki / Hamachan
> >
> >
> >
>
------------------------------
Date: Thu, 8 May 2003 12:00:58 -0500
From: Joseph Gathman <jpgathman@MSN.COM>
Subject: adding variances - related question...
>Doug wrote: "I've put this in the category of "things I know where to lo
k
>up when I need them, but not to carry in my head" Perhaps that's wrong.
Well, if you are wrong, so am I. And here is another one for that category:
what to do with the SD or SE when scaling density measures to a standard,
such as the conversion of # benthic invertebrates per sample (of specific,
but inconvenient, benthic surface area) to # of critters per square meter.
I always have to stop and think about the correct way to do it, but I know
some people don't even realize the need for a conversion. I've seen it done
wrong even in published papers.
Joe Gathman
_________________________________________________________________
Add photos to your e-mail with MSN 8. Get 2 months FREE*.
http://join.msn.com/?page=features/featuredemail
------------------------------
Date: Thu, 8 May 2003 15:13:09 -0400
From: Lori Hidinger <lori@ESA.ORG>
Subject: Towards Best Practices -- An On-line Forum
Towards Best Practices (TBP) is a new electronic forum, built as a
resource for the conservation community. The eForum, at
www.nbii.gov/datainfo/bestpractices/index.html allows researchers,
policy-makers, students, and others to develop principles and practices
related to the study and management of biocomplexity and conservation.
Papers published on this new eForum will be open for review and
discussion in a free, public electronic community. Entries will also be
eligible for one of several Awards for Excellence, including sponsorship
of travel to the annual meeting of the Society for Conservation Biology.
Go to www.nbii.gov/datainfo/bestpractices/index.html for more
information about submitting papers, starting discussions, and about the
rich range of resources on conservation practices available through the
eForum. At this site, you can also find more information about the
Awards for Excellence in Best Practices available to all who are
interested the study or management of the complex interactions between
life forms - including human populations - and the environment.
------------------------------
Date: Thu, 8 May 2003 10:57:46 -0500
From: Chris Brown <cabrown@TNTECH.EDU>
Subject: posting of pdf versions of manuscripts on websites/ stats discussio
To all,
A comment on the stats discussion, then a question on an unrelated
topic.
First, one thing I think being somewhat lost in the discussion is
the difference between (1) needing to recognize that combining data with
associated errors is more complex than just adding the means together,
and (2) knowing the exact way to do this. If I was teaching a stats
class, I'd be extremely happy if the students knew (1) without knowing
exactly how to go about figuring out the right answer off the top of
their head, rather than knowing the formula for (2) without having any
idea about why or when it should be used. And the latter often seems to
be the case...
Second, I hope to (finally, as far as my students are concerned!)
get my web site up and running this summer, and have a question I've not
seen addressed before. What, if any, are the legalities associated with
posting pdf copies of my published reprints on the web? In other words,
am I free to simply scan in reprints and have a link to them, treating
this as no different than someone mailing me and asking for a reprint?
Or must I get approval from the journals...who are, after all, copyright
holders and would charge $10 or more to access the same article I'm
providing for free. Any thoughts will be most appreaciated.
Chris
*********************************************
Chris Brown, PhD
Assistant Professor
Department of Biology
Box 5063
Tennessee Technological University
Cookeville, TN 38505
phone: (931) 372-6258
email: cabrown@tntech.edu
------------------------------
Date: Thu, 8 May 2003 09:54:08 -0700
From: Bill Grantham <grantham_bill@HOTMAIL.COM>
Subject: Re: US to open public wilderness lands for...
<html><div style='background-color:'><DIV></DIV>
<P>Hi All,</P>
<P>I was wondering if anyone out there could provide some clarificatio
and perspective on what this decision means. </P>
<P>The article states the obvious - "By declaring an end to wild land
urveys, the administration ruled out protection of these areas as formal wil
erness." and "Now, these areas, managed by the BLM, could be re-opened to mi
ing, drilling, logging, and road-building."</P>
<P>With the cessation of the reviews of areas for new wilderness prote
tion, does this also mean a drastic reduction in research for locating and m
nitoring of rare plant and animal populations on federal lands? Does this me
n a drastic reduction of future Federal funding of research after current gr
nts run out? </P>
<P>Finally, what else can we do in addition to writing our congress pe
ple and voting in 2004?</P>
<P>Bill Grantham</P>
<P>Graduate Research Assistant</P>
<P>West Virginia University<BR><BR></P>
<DIV></DIV></div><br clear=all><hr>Protect you
PC - <a href="http://g.msn.com/8HMEENUS/2755??PS=">Click here</a>
for McAfee.com VirusScan Online </html>
------------------------------
Date: Thu, 8 May 2003 08:33:36 -0700
From: Patrick Foley <patfoley@CSUS.EDU>
Subject: Re: Statistics teaching
Ecologists,
To clarify my previous post:
The standard error of an estimate (say a mean) is the standard deviation
of the values the mean would take if we redid the sample numerous times.
There is nothing special about a standard error; it is just a standard
deviation (but of the mean not the original data). Other kinds of
estimates (besides means) can also be assigned standard errors in the
same way.
So if you want to estimate Z = X +Y, and you have estimates for X and Y
and their standard errors, then the expected value of Z is EZ = EX + EY,
since expected values E are linear operators. And VZ=VX+VY+2Cov(X,Y).
Usually we assume that X and Y are roughly independent, in which case
the covariance term drops out.
"Error propagation" is just a case of simple probability rules applied
to the sums of random variables.
Patrick Foley
patfoley@csus.edu
Patrick Foley wrote:
>If two random variables X and Y are independent, then their variances
>are additive. So first convert the standard errors to "standard
>variances" (not a standard term) by squaring them. Then add them. Then
>take the square root to get the cumulative standard error.
>se(X+Y)=sqrt(se(X)^2+se(Y)^2)
>
>If X and Y are correlated, you must add in a covariance term.
>
>Patrick Foley
>patfoley@csus.edu
>
>David C Baker wrote:
>
>
>
>>Greetings, all.
>>I am almost afraid to respond here to Bill Silvert's little riddle.
>>Intuitively one might say that 10(+/-3) snails/m2 and 20(+/-4)
>>earthworms/m2 add up to 30 (+/-7) inverts. But something about how
>>standard errors are calculated make me skeptical of that answer. I
on't
>>have a basic statistics book in front of me, and I've forgotten the
ormula
>>in the past few years. Maybe some of the much more savvy than I out
there
>>can give me the easy answer: is it purely additive? Seriously, with
tongue
>>firmly in cheek.
>>
>>regards
>>David Baker
>>Central Oregon Ecology Program
>>(541) 383-5424
>>
>>
>>
>>
>>
>
>
>
------------------------------
Date: Thu, 8 May 2003 18:28:03 +0100
From: William Silvert <silvert@NETCABO.PT>
Subject: Combining standard errors
Now that the wave of comment and conjecture has died down, yes, it is indeed
the sum of squares - since 3^2+4^2=5^2, the sum of 10+_3 and 20+_4 is 30+_5.
For multiplication the result is similar, except that it is relative
errors - (100+_3) x (100+_4) = 10000+_500 (errors of 3% and 4% give 5%).
For subtraction and division the results are the same as for addition and
multiplication.
Someone suggested that this is just a technical detail, a sort of
grammatical nicety. I disagree, and feel that every scientific value should
have a meaningful confidence limit associated with it. Physicists often
laugh at those of us who produce imprecise numbers like the biomass of a
forest to seven significant figures, but I think that a lack of awareness of
how meaningful the numbers we produce are likely to be is a serious
shortcoming. And yes, there are cases where meaningless numbers have been
produced because of a lack of attention to confidence limits.
Bil Silvert
------------------------------
Date: Thu, 8 May 2003 14:08:57 -0400
From: Resit Akcakaya <resit@RAMAS.COM>
Subject: Re: [Ecolog] citation for Var(X-bar*Y-bar) {01}
Hi Mike,
I believe your formula might be wrong; I think the final term should
be added, not subtracted. Equation (2) in Goodman (1960) is
V(xy) = X^2 V(y) + Y^2 V(x) + V(x) V(y)
X (uppercase) is defined as E(x), i.e., the mean of the random
variable x, and V(x) is defined as its variance (the definitions are
in the paragraph preceding equation 1).
What might be confusing is equation (5) on the following page, which
has a negative third term:
v(xy) = x^2 v(y) + y^2 v(x) - v(x) v(y)
(note that all these are lowercase). Here, x (lowercase) is a
random variable (not its mean).
The equation you need to calculate the variance of a product (based
on means and variances of two random variables) is eq(2); thus, I
believe the last term (the product of variances) should be added,
not subtracted.
Also, note that this formula assumes that the variables x and y are
independent; equation 18 gives the general formula, which is rather
complex (eq 19 gives an approximation).
Resit Akcakaya
----------------------------------------------
Applied Biomathematics
100 North Country Road, Setauket, NY 11733, USA
Tel: 631-751-4350 Fax: 631-751-3435
resit@ramas.com http://www.ramas.com/resit.htm
Larson, Michael Andrew wrote:
> Not long ago I had difficulty finding the proper way to calculate the v
riance of a product; I was calculating the fecundity terms (fertility*surviv
l) for a population projection (i.e., Leslie) matrix. As some have mentione
earlier on this thread, many basic stats texts do not provide such equation
for propagation of error. The lone exception seems to be that most of the
exts I have provide guidance for the SE of a difference between two means.
>
> For those who may be interested,
>
> Var(X*Y) = [X^2 * Var(Y)] + [Y^2 * Var(X)] - [Var(X) * Var(Y)]
>
> where Var=variance, X=mean of sample 1, Y=mean of sample 2, and X^2=X-s
uared.
>
> Please check the original publication in case I copied something incorr
ctly:
> Goodman 1960 J. Amer. Stat. Assoc. 55:708-713.
>
> Mike Larson
> Fisheries & Wildlife
> University of Missouri
>
------------------------------
Date: Thu, 8 May 2003 22:00:44 +0100
From: William Silvert <silvert@NETCABO.PT>
Subject: Two unrelated answers
With regard to Chris' first point, I really don't worry about the exact way
to estimate error, which seems to me like an oxymoron. I have seen rigorous
derivations of the results I gave, but if I am off by a few percent in the
size of the confidence region, I don't think that matters very much - the
important point is being sure to calculate at least an approximate
confidence region.
As for the second point, I avoid posting exact copies of copyrighted
publications, but generally I post a draft - like a preprint. It contains
almost the same material, but because I usually go from word processor
format to HTML rather than posting the final PDF file, it is not a perfect
copy, and it is not the final version anyway. I have never had complaints
about posting preprints.
I think that something like this should be done, the number of people
without access to expensive journals is exploding.
Bill Silvert
----- Original Message -----
From: "Chris Brown" <cabrown@TNTECH.EDU>
To: <ECOLOG-L@LISTSERV.UMD.EDU>
Sent: Thursday, May 08, 2003 4:57 PM
Subject: posting of pdf versions of manuscripts on websites/ stats
discussion
> To all,
> A comment on the stats discussion, then a question on an unrelated
> topic.
> First, one thing I think being somewhat lost in the discussion is
> the difference between (1) needing to recognize that combining data wit
> associated errors is more complex than just adding the means together,
> and (2) knowing the exact way to do this. If I was teaching a stats
> class, I'd be extremely happy if the students knew (1) without knowing
> exactly how to go about figuring out the right answer off the top of
> their head, rather than knowing the formula for (2) without having any
> idea about why or when it should be used. And the latter often seems to
> be the case...
> Second, I hope to (finally, as far as my students are concerned!)
> get my web site up and running this summer, and have a question I've no
> seen addressed before. What, if any, are the legalities associated with
> posting pdf copies of my published reprints on the web? In other words,
> am I free to simply scan in reprints and have a link to them, treating
> this as no different than someone mailing me and asking for a reprint?
> Or must I get approval from the journals...who are, after all, copyrigh
> holders and would charge $10 or more to access the same article I'm
> providing for free. Any thoughts will be most appreaciated.
>
> Chris
> *********************************************
> Chris Brown, PhD
> Assistant Professor
> Department of Biology
> Box 5063
> Tennessee Technological University
> Cookeville, TN 38505
> phone: (931) 372-6258
> email: cabrown@tntech.edu
>
------------------------------
Date: Thu, 8 May 2003 16:08:41 -0500
From: Dave McNeely <dlmcneely@LUNET.EDU>
Subject: Re: posting of pdf versions of manuscripts on websites/ stats
discussion
----- Original Message -----
From: "Chris Brown" <cabrown@TNTECH.EDU>
To: <ECOLOG-L@LISTSERV.UMD.EDU>
Sent: Thursday, May 08, 2003 10:57 AM
Subject: posting of pdf versions of manuscripts on websites/ stats
discussion
What, if any, are the legalities associated with
> posting pdf copies of my published reprints on the web? In other words,
> am I free to simply scan in reprints and have a link to them, treating
> this as no different than someone mailing me and asking for a reprint?
> Or must I get approval from the journals...who are, after all, copyrigh
> holders and would charge $10 or more to access the same article I'm
> providing for free. Any thoughts will be most appreaciated.
I know that a lot of people do it. Dave McNeely
------------------------------
Date: Thu, 8 May 2003 16:47:08 -0500
From: "Lemke, Michael" <Lemke.Michael@UIS.EDU>
Subject: Aquatic Ecology Graduate Assistantship, Univ. of IL-Springfield
Graduate Research Assistantship (M.S.): Department of Biology, University
of Illinois at Springfield. Focus: aquatic ecology of Illinois River
floodplain.
A graduate research assistantship is available for a highly motivated
student to investigate nutrient dynamics and microbial ecology of the
Illinois River floodplain. The successful candidate will be able to
collaborate with other graduate students and P.I.'s engaged in the study of
floodplain ecology. This position can begin as early as June with 2 years
of support (tuition and stipend) at the University of Illinois at
Springfield. Applicants must possess a B.S. or B.A. in Biology or
equivalent, be admitted to the Biology M.S. program
(http://www.uis.edu/~biology), and possess a valid driver's license.
Preferred applicants should demonstrate abilities &/or willingness to sample
in various weather conditions, conduct experiments, process samples, analyze
data, and write effectively. Experience in microbial ecology (molecular
emphasis), aquatic ecology, or water chemistry are helpful, but not
required. Applications should include: 1) letter of interest, 2) resume,
and 3) names, addresses, phone numbers and e-mail addresses of three
references. Applications, or request for more information, may be sent to:
lemke.michael@uis.edu (email is preferred) - or - Dr. Mike Lemke,
University of Illinois at Springfield, Biology Department, One University
Plaza, Springfield, IL 62703. Applications accepted immediately; position
open until filled by qualified applicant.
------------------------------
Date: Thu, 8 May 2003 17:37:43 -0700
From: David Bryant <dmb@IO.HARVARD.EDU>
Subject: Re: Combining standard errors
It is very gratifying to see this much productive discussion on a topic
which, I agree with Bill et al, that this deserves more attention in
ecological data analysis. Now that we've dusted off those tired neurons in
which we kept our stats knowledge, I have a practical application for the
topic which transcends the simple combination of means.
I am uncomfortable with the lack of TRUE error statements ( I prefer the
less pejorative "variance") in ecological literature. While this may not
pertain to organismic researchers, community and ecosystem level ecologists
and biogeochemists are often guilty of stating only a fraction of the true
error in their data. I consider myself in all 3 of these categories and
sheepishly admit to casting the first stone. As an example , consider that
one is testing the hypothesis that leaf litter quality relates to soil
invertebrate density, or soil CO2 flux etc. First collect some litter and
measure some chemical constituents, say N and Lignin content. Already we
have multiple error terms about these numbers: the variance associated with
field sampling, variance of the subsampling for chemical analysis, the
machine (or wet chemistry) precision and the variance among the
replicates. And we haven't even started measuring the independent
variable! Now imagine how this propagates when combining various
reservoirs, flux rates and residence times when calculating elemental
budgets! Is your head swimming yet?
How often are these errors combined in the data presented for publication?
Of course we could write programs to combine these error terms, now that we
have several copies of the equation. But does stats software exists that
would do this for us?
While such error reckoning may broaden our error bars and reduce the
statistical power of our analyses I fell that it would improve the ability
to compare published data in meta-analysis and advance our scientific
credibility in the long run.
David
David M. Bryant Ph. D.
Dept of Earth and Planetary Sciences
Harvard University
20 Oxford St.
Cambridge, MA 02138
dmb@io.harvard.edu
617-496-6246
------------------------------
Date: Thu, 8 May 2003 16:57:17 -0400
From: Steve Friedman <friedm69@MSU.EDU>
Subject: Re: US to open public wilderness lands ..
<html>
The first thing we should plan on doing is <b><u>elect new publi
officials </u></b>in the next series of state and federal
elections. <br>
<br>
Secondly, for those of us who remember James Watt, didn't he try to do
this as well?<br>
<br>
Third, keep fighting the good fight. Write to your congress person
and express outrage at this decision. <br>
<br>
Fourth, join conservation organizations, pay dues and help those
organization lobby. When environmental planning has been ignored by
past administrations the membership of environmental organizations tend
to swell. It is important to maintain this trend. <br>
<br>
I'm sure there are many more actions that we can take, I just can't think
of more right now.<br>
<br>
Steve Friedman<br>
<br>
<br>
<x-sigsep><p></x-sigsep>
----------------------------------------------------------------------------
-----------------------------------------<br>
Steve Friedman<br>
<br>
Assistant Professor Forest Management / GIS<br>
Departments of Forestry & Geography<br>
126 Natural Resources<br>
Michigan State
University<x-tab> </x-tab>
<x-tab> </x-tab><
x-tab> </x-tab><b
>
East Lansing, Michigan 48824<br>
<br>
Office:<x-tab> </x-tab>517 - 353 - 9230<br>
Fax:<x-tab> </x-tab><x-tab> &n
sp; </x-tab>517
- 432 - 1143<br>
email
<x-tab> </x-tab><x-tab>
nbsp; </x-tab>friedm69@msu.edu</html>
------------------------------
Date: Thu, 8 May 2003 15:34:27 -0700
From: "Jeffrey D. Corbin" <corbin@SOCRATES.BERKELEY.EDU>
Subject: Re: US to open public wilderness lands ..
Speaking of lobbying, conservation organizations that accept tax-deductible
contributions CANNOT lobby for political purposes. So, if you want to have
maximum impact on legislative decisions, consider contributing to
organizations such as The Sierra Club that can directly contribute to
campaigns, etc. The absence of an environmental lobby in Washington and
elsewhere, especially when compared to the Industry Lobby, is a major
impediment to getting a seat at the table when decisions are made.
Jeff Corbin
>Fourth, join conservation organizations, pay dues and help those
>organization lobby. When environmental planning has been ignored by pas
>administrations the membership of environmental organizations tend to
>swell. It is important to maintain this trend.
>
>I'm sure there are many more actions that we can take, I just can't thin
>of more right now.
>
>Steve Friedman
>
>
>------------------------------------------------------------------------
---------------------------------------------
>Steve Friedman
>
>Assistant Professor Forest Management / GIS
>Departments of Forestry & Geography
>126 Natural Resources
>Michigan State University
>East Lansing, Michigan 48824
>
>Office: 517 - 353 - 9230
>Fax: 517 - 432 - 1143
>email friedm69@msu.edu
****************************************
Jeffrey D. Corbin
Department of Integrative Biology, VLSB#3060
University of California
Berkeley, CA 94720-3140
(510) 643-5430
http://socrates.berkeley.edu/~corbin
****************************************
------------------------------
Date: Thu, 8 May 2003 16:53:08 -0600
From: Dave Whitacre <dwhitacre@PEREGRINEFUND.ORG>
Subject: ECOLOG--posting PDF files to web
I don't know the legalities. However I recently contacted a couple =
journal editors about posting my reprints to the web as PDFs. One editor =
(a well-known bird journal) said "Feel free to, as long as I'm editor."
The other editor (of a museum bulletin) was not at all thrilled at the =
prospect (which involved a 100 pg monograph), and asked me not to post =
it.
In this latter case, it seems plausible that such a posting could amount =
to serious competition for sales of a separately issued volume of a =
bulletin with modest circulation.
David Whitacre
The Peregrine Fund
5668 West Flying Hawk Lane
Boise, Idaho 83709
(208) 362-3716
------------------------------
Date: Thu, 8 May 2003 15:42:24 -0700
From: Theodore Kennedy <tkennedy@USGS.GOV>
Subject: Re: posting of pdf versions of manuscripts on websites/ stats
discussion
I believe posting pdfs on a personal website for others to download is
legal, as is sending reprints through the mail. However, sending someone
an email with a article attached in pdf or other format is technically
illegal.
Don't ask me why......
tk
please note new phone and address
****************************************************************************
Ted Kennedy
Grand Canyon Monitoring & Research Center
USGS
2255 N. Gemini Dr.
Flagstaff, AZ 86001
(928) 556-7322
tkennedy@usgs.gov
|---------+---------------------------->
| | Dave McNeely |
| | <dlmcneely@lunet.|
| | edu> |
| | Sent by: |
| | "Ecological |
| | Society of |
| | America: grants, |
| | jobs, news" |
| | <ECOLOG-L@LISTSER|
| | V.UMD.EDU> |
| | |
| | |
| | 05/08/2003 02:08 |
| | PM |
| | Please respond to|
| | Dave McNeely |
| | |
|---------+---------------------------->
>----------------------------------------------------------------------
-------------------------------------------------------------------|
|
|
| To: ECOLOG-L@LISTSERV.UMD.EDU
|
| cc:
|
| Subject: Re: posting of pdf versions of manuscripts on websites/
tats discussion |
>----------------------------------------------------------------------
-------------------------------------------------------------------|
----- Original Message -----
From: "Chris Brown" <cabrown@TNTECH.EDU>
To: <ECOLOG-L@LISTSERV.UMD.EDU>
Sent: Thursday, May 08, 2003 10:57 AM
Subject: posting of pdf versions of manuscripts on websites/ stats
discussion
What, if any, are the legalities associated with
> posting pdf copies of my published reprints on the web? In other words,
> am I free to simply scan in reprints and have a link to them, treating
> this as no different than someone mailing me and asking for a reprint?
> Or must I get approval from the journals...who are, after all, copyrigh
> holders and would charge $10 or more to access the same article I'm
> providing for free. Any thoughts will be most appreaciated.
I know that a lot of people do it. Dave McNeely
------------------------------
Date: Thu, 8 May 2003 20:55:35 -0400
From: "Luis J. Villanueva-Rivera" <ejasperi@YAHOO.COM>
Subject: Re: posting of pdf versions of manuscripts on websites
Hi,
ESA has a statement on "Permissions for reprinting and republishing" of ESA
journals articles (a portion of it appears below) which can serve as an
example of what some journals will allow. I think each author should check
with the journal, but my feeling is that most journals will have no problem
if the author post their articles in their personal webpage.
"Authors may post their articles to their personal or home institution's
website and may make and distribute photocopies of such articles.
Instructors may post up to 10 articles from ESA publications to a publicly
available class website, or an unlimited number of articles to a website or
other electronic medium with access restricted to their class. To copy or
transmit otherwise, to republish, to post on public servers, to redistribute
to lists, or to use any component of an article from an ESA journal in other
works, or to use such an article for commercial or promotional purposes,
requires prior specific permission and possible fee."
The full document can be found at
http://www.esapubs.org/esapubs/permissions_main.htm
Luis J. Villanueva-Rivera, B.S.
San Juan, Puerto Rico
ejasperi@yahoo.com
http://www.CoquiPR.com
------------------------------
End of ECOLOG-L Digest - 7 May 2003 to 8 May 2003 (#2003-122)
*************************************************************
˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙
Thanks to discussion with TVR, I have decided to put a link to back files of the discussion group. This months back files.
The link to complete archives is available elsewhere.
This text was originally an e-mail. It was converted using a program
RUPANTAR- a simple e-mail-to-html converter.
(c)Kolatkar Milind. kmilind@ces.iisc.ernet.in