Re: 關於建設新臺灣國語 ~ a statistical and practical initial assessment now is open for comment.

From: bingh tang
Date: 2010-01-15 01:00:32

For the time being my comment  on the statistical experiment is as follwoing, but sorry that I can't translate all of them into Taigu or Sino at thie moment :

   

My concern is that the actual point of comparison is cross-over vs. parallel groups.

-the question of "superiority of one language to the other " is logically a potentially false one; many contexts simply cannot support one or the other.

1] -in the instance that both are appropriate, the late John Fleiss gives some kind of estimate of efficiency -the power that can be obtained will require some percentage of fewer participants in a cross over study, relative to parallel groups, given that ALL other parameters are held constant in the problem.

 

2]

so: there are several reasons why this is not reasonable; if all are held constant then Lin hianN is correct that the variance being limited increases the chances of a statistically significant signal - but says nothing about the practical significance of the resulting signal. [ 青盲仔毋驚蛇] 相爭一跤箸，放去一隻羊!! ~ 因小失大。