To All, I receive TiPS in digest form so I will respond to 2 messages at that same time as well as provide one admonishment.
----- Original Message ----- From: "Teaching in the Psychological Sciences digest" <[EMAIL PROTECTED]> To: "tips digest recipients" <[EMAIL PROTECTED]> Sent: Sunday, September 19, 2004 12:00 AM Subject: tips digest: September 18, 2004 > TIPS Digest for Saturday, September 18, 2004. > > 1. Re: International poll on US election > 2. Re: stat question > 3. Re: stat question > 4. [Fwd: Re: [Fwd: Re: Representative Samples and the 2004 Vote]] > > ---------------------------------------------------------------------- > > Subject: Re: International poll on US election > From: "Christopher D. Green" <[EMAIL PROTECTED]> > Date: Sat, 18 Sep 2004 09:53:48 -0400 > X-Message-Number: 1 To Prof. Green: I don't know why but your messages are followed by a copy in html. Could you check to see if this can be turned off in the interest of saving bandwidth and reducing message size? Original message: > This is a multi-part message in MIME format. > --------------040505080609030309040807 > Content-Type: text/plain; charset=us-ascii; format=flowed > Content-Transfer-Encoding: 7bit > [snip] > > --------------040505080609030309040807 > Content-Type: text/html; charset=us-ascii > Content-Transfer-Encoding: 7bit Green's duplicate message: > <!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"> > <html> > <head> > <meta http-equiv="Content-Type" content="text/html;charset=ISO-8859-1"> > <title></title> > </head> > <body text="#000000" bgcolor="#ffffff"> > Allen Esterson wrote:<br> > <blockquote type="cite" > cite="[EMAIL PROTECTED] stburg.edu">Rick > wrote [snip]<br> > <blockquote type="cite">For the record, George Washington, and > several other founding fathers of<br> >[snip] > <p></p> > </div> > --- <BR> You are currently subscribed to tips as: [EMAIL PROTECTED] <BR> To unsubscribe send a blank email to [EMAIL PROTECTED] </BODY> > </html> > > --------------040505080609030309040807-- > > > ---------------------------------------------------------------------- > > Subject: Re: stat question > From: "Stephen Black" <[EMAIL PROTECTED]> > Date: Sat, 18 Sep 2004 11:56:04 -0500 > X-Message-Number: 2 > > On 17 Sep 2004, Martin J. Bourgeois wrote: > > > I know that this is a little off-topic, but it could be > > teaching-related for stat instructors. Does anyone know how to > > test whether or not the difference between two ranges (in a > > between-subjects design) is statistically significant? An editor > > asked me to do this test, and I have never heard of such a test, > > nor can I find one. I've lost my copy of Dixon and Massey's "Introduction to Statistical Analysis" (you library may well have a copy) but I do remember that they provided a variety of "non-traditional" tests involving ranges. However, I really must ask about the reasoning for why a comparison of ranges was requested, because (a) as estimates of a population variability (actually, variance -- if you assume that the sample was drawn from a normal distribution which ranges from negative infinity to positive infinity, the the range is not a population parameter, rather it is an indicator of the popuplation's variance), the range is extremely unstable and the sampling distribution of ranges will be much wider than that of the same variance, and (b) the statistical power for a test of differences between ranges (which I believe would be a ratio of the larger range to the smaller range, like the F test for testing equality of variances) is likely to be quite low unless you have some ridiculuously large sample size (which again raises the question as to why one would do such a test instead of a direct test on variances). If Dixon & Massey doesn't have the test, then I'd suggest asking a local statistician on how one might perform a permutation test on the ranges (comparable to a Fisher's exact test on a 2x2 table but involving the ranges -- See Efron & Tibshirani's "An Introduction to the Bootstrap", Chapter 16, for a general introduction). > ---------------------------------------------------------------------- > > Subject: [Fwd: Re: [Fwd: Re: Representative Samples and the 2004 Vote]] > From: "Christopher D. Green" <[EMAIL PROTECTED]> > Date: Sat, 18 Sep 2004 12:48:11 -0400 > X-Message-Number: 4 > > This is a multi-part message in MIME format. > --------------080702090808040207040602 > Content-Type: text/plain; charset=us-ascii; format=flowed > Content-Transfer-Encoding: 7bit > > Since I'm already in trouble for broaching "political" topics here (use > your delete key now!), here is one which bears greatly on topics of fair > statistical analysis of data (psychological or otherwise). What is > George Gallup's political affiliation anyway? Read on... > Christopher Green > > -------- Original Message -------- > http://www.theleftcoaster.com/archives/002806.html > > ==== > > Why You Should Ignore The Gallup Poll This Morning - And Maybe All Of Theirs [Snip article which ultimately points out that Gallups CEO is a GOP contributor] For those interested in such things, I'd suggest looking at the website www.electoral-vote.com a website that attempts to describe and projects the electoral college's vote (which actually determines who is become the U.S. President). The "Votemaster" provides access to a variety of data as well as providing interesting IMHO daily commentary on how/why the poll results are/might be what they are. The following is the duplicate of Prof. Green's message in html. > --------------080702090808040207040602 > Content-Type: text/html; charset=us-ascii > Content-Transfer-Encoding: 7bit > > <!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"> > <html> > <head> > <meta http-equiv="Content-Type" content="text/html;charset=ISO-8859-1"> > </div> > --- <BR> You are currently subscribed to tips as: [EMAIL PROTECTED] <BR> To unsubscribe send a blank email to [EMAIL PROTECTED] </BODY> > </html> > > --------------080702090808040207040602-- Michael Palij [EMAIL PROTECTED]
