The following is an archived copy of a message sent to a Discussion List run by the Campaign Against Sanctions on Iraq.
Views expressed in this archived message are those of the author, not of the Campaign Against Sanctions on Iraq.
[Main archive index/search] [List information] [Campaign Against Sanctions on Iraq Homepage]
--- Darin Zeilweger <email@example.com> wrote: > > Don't read too much into those polls, now. > > Asking 1,011 people a question and then > extrapolating that number across > a population of almost 300,000,000 is dubious at > best. <Begin crash course in basic statistics> That depends on what conclusion you want to draw from it. Assuming a fair poll, if your conclusions are based on "7 out of 10 people" then you are stating that between 65 and 75% of people..... Assuming that they got a result of 70% of people..... The data is binomially distributed. Ha: The probability that any person agrees is >=0.65 H0: The probability that any person agrees is < 0.65 Now take a test where 1011 people are asked and 70% agree (that means ~707 people agree) By my calculation: p(x<=657)=0.99957 That's one hell of a confidence! It's "not totally, absolutely, completely, 100% sure - mathematically" at worst. <end crash course in basic statistics> Alun Harford __________________________________ Do you Yahoo!? Yahoo! SiteBuilder - Free, easy-to-use web site design software http://sitebuilder.yahoo.com _______________________________________________ Sent via the discussion list of the Campaign Against Sanctions on Iraq. To unsubscribe, visit http://lists.casi.org.uk/mailman/listinfo/casi-discuss To contact the list manager, email firstname.lastname@example.org All postings are archived on CASI's website: http://www.casi.org.uk