Only an 'F', hey! Your wife must be a hard marker, lucky I'm not in her class :)
Have a look at http://whyfiles.org/009poll/math_primer.html for the phoney formula And http://www.amstat.org/sections/srms/brochures/margin.pdf pages 4,5,6, 7 and 8 in particular if you are interested. Graeme got the long winded version direct as I am sure you lot have had enough of statistics! (well at least 73% of you with a margin of error of 6%) Cheers David (who enjoyed some very pleasant ridge flying on the weekend, thinking about statistics not at all, until the largest red back spider I ever saw appeared from under the instrument panel cover, then disappeared again. Kinda puts you off the rest of your flight!) >-----Original Message----- >From: [EMAIL PROTECTED] >[mailto:[EMAIL PROTECTED]] On Behalf Of >Graeme Cant >Sent: Sunday, 23 June 2002 11:56 PM >To: [EMAIL PROTECTED] >Subject: RE: [aus-soaring] RPL results > > >>From: "David Conway" <[EMAIL PROTECTED]> >>Don't call the kettle black (if you are a pot)! > >You picked me, David! My wife's the stats teacher, not me. > >She gives you an 'F' for the phoney error calculation. :) > >As you say at the bottom, the sample must be representative of the >population (only THEN would the population size not matter). With a >self-selected sample, the only chance of useful information >(well below the >level of valid stastistical inference) is to have a very large >sample. >Ron's was tiny. > >That's why the GFA didn't amalgamate. With a voluntary vote, >they still >couldn't be sure about the opinion of the non-voting members. In your >stats, they would have been justified in amalgamating because >the probable >'error' was just over 4%. > >A self-selected sample is a biased sample and an 'error' >calculation has no >meaning. No valid inferences can be made from a bodgy sample >about those >not in the sample. You can make no valid inferences about all >the pilots >who DIDN'T vote (either for amalgamation or for an RPL). > >> >I think rational discussion is a better basis for decision making. > >I still prefer that to misleading statistics. :) > >Graeme Cant > > -- * You are subscribed to the aus-soaring mailing list. * To Unsubscribe: send email to [EMAIL PROTECTED] * with "unsubscribe aus-soaring" in the body of the message * or with "help" in the body of the message for more information.
