Hi Opaqueice, I am a psychologist or (student of such subject) ;) so I am prity into statistics and signifficance as layed out by Ronald Fisher. Based on the works of Bernuille he developed a test methodology to determine if a certain finding is more than accident i.e has a systematic cause. Neyman and Pearson further developed the concept and eventually ended up with what statisticians call "sequential testing". Some people say it makes a difference if you put in the tea after the milk or vice versa. Other doubt that. If now a Lady comes along and says, well I will determine surely what you have put in to my cup of tea we need to find a schema of how to determine if this is just guess or really the Lady having some sort of idea how her tea was made. To be sure we will repeat our tea meetings every sunday afternoon. Now how many afternoons will we sit until we finally believe her? That will obviously depend on a regorous testing and on the number of right and wrong guesses. Every sunday she makes a decision and after ten sundays she has 1024 possible paths of guesses. First right, second right, third wrong... Or first right, second wrong third wrong, fourth right. These 1024 paths fall into ten categories. Ten right. or nine right one wrong. or eight right and so forth. Obviously there are more paths that lead to a fifty fifty result. I.e 5 right and 5 wrong. There are actually 252 such possibilities. However there is only one possible course that returns ten out of ten wrongs or rights. Apparently the more extreme we get the less likely it is that it is an accident. By saying that I reached signifficant recognition rates of .0531% I mean that there is only 1024*0.0531 56 approx. (had calculated with 55) possibilities of have 8 out of 10. "Sequential" now means that the number of test repetitions is not predetermined but that the test is carried out until error likelyhood drops under a predetermined amount which normally is about 5% in published studies of any matter. Of course one can be more demanding and only publish ones findings when the error likelyhood drops under 1%. If we take our 1024 paths out of ten sundays or songs listened to this would make for only 10,24 paths with wrong answered in it. This is the case either if we have no wrongs or if we just have one wrong which can of course occur on the first or second or third or last occation. I listened to whole songs and then gave my judgement. When I say "signifficantly" different I just mean statistic signifficance as described not subjective or even accustic signifficance for the first is not mesurable and the second would be the subject of a detailed wave analysis which I think is out of the scheme of this undertaking ;) Cheers Jeronimo PS.: Enjoy your tea. Which one came first? Milk or tea? PPS.: The english say that there is truly a difference caused by already resident milk slowly warmed up by flooding tea compared to a hot cup of tea which drastically warms up the milk which is being added.
-- jaysung ------------------------------------------------------------------------ jaysung's Profile: http://forums.slimdevices.com/member.php?userid=12375 View this thread: http://forums.slimdevices.com/showthread.php?t=38653 _______________________________________________ audiophiles mailing list [email protected] http://lists.slimdevices.com/lists/listinfo/audiophiles
