|
test From: welcome.perfect@mac.com Subject: Automation of statistical analysis of ustream and You Tube Date: August 5, 2008 5:40:31 AM PDT To: support@ustream.tv, cancel@ustream.tv, Press@ustream.tv, bizdev@ustream.tv, sales@ustream.tv, webserver@ustream.tv Cc: Press@ustream.tv, bulletin@law.harvard.edu 2008:08:05:08:24:75*W:Null Hyothesis: This is what my friends are concerned about. "People with weapons of mass destruction should live in glass houses," is one response. Statistical tools to randomly sample date:times from a calendar and then test what hypotheses is another response. For example between June 21st, 2008 and August 5th, 2008 there are x number of days expressed as minutes ['Whump]. Random numbers between 000 and 100 permit a sample of minutes to be selected from the chronologically ordered sample space. [Whump] Then the hypothesis, say, "All work, and no play, makes Jack a dull boy," might be tested, although the fact Jack may have had a lot of work, or play outside the sample space could be a significant factor. Or, say, "Repentance is not a significant learning strategy in whose life," Then the samples might show the word 'repentance' comes up often enough to be significant, although there's the experimental design issue of can one occurrence be SO significant you are unlikely to be able measure significance by frequency? How about, ['whn] "Who is self directed", or "Who is other directed", or "Who is super-ego oriented", or "Who is 'id' oriented"? In any <b>...</b> |
Views:
554
0
ratings |