On Sun, Oct 19, 2008 at 4:24 PM, William T Goodall <[EMAIL PROTECTED]>wrote:
> > On 19 Oct 2008, at 22:48, Nick Arnett wrote: > > During the weeks following my post, we had abnormally high activity > > -- more > > than two standard deviations from the mean. One of those weeks, we > > were at > > more than five standard deviations above the mean. > > Obviously controversy always causes a flurry of postings. For a true > picture one must wait until the list quiesces again. How... convenient for you to imagine that the data isn't meaningful when it contradicts you. Running for office, are you? Anybody else think that explains six weeks of high posting activity followed by two weeks of normal activity, very little of which had anything to do with the post William would have us believe discouraged activity? William, you actually haven't documented a statistically significant decrease in postings (because there isn't one). Maybe it's time for a review of basic statistics? The reason that the last two weeks' posting level doesn't represent a significant drop is that 90 percent of values in a normal distribution fall within one standard deviation of the mean. The "decline," in terms of statistical significance, is a pink unicorn. Shoo it away. I have started two companies and have four patents for this kind of analysis and eight years of experience doing it. My patents were bought by Nielsen/BuzzMetrics and my second startup was acquired by one of the biggest social media agencies in the world. So, what do you have to back up your analysis? Like, you know, numbers and stuff? Or, and this is my preference, just drop it. You are completely wrong; the numbers don't lie. Nick _______________________________________________ http://www.mccmedia.com/mailman/listinfo/brin-l
