On 4 Mar 2004 06:04:41 -0800, [EMAIL PROTECTED] (Will) wrote: > I don't understand this. Can anyone say this in another way, perhaps a > more mathematical way? > > "Why stronger relations between variables are more significant. > Assuming that there is no relation between the respective variables in > the population, the most likely outcome would be also finding no
I think you need to absorb that the comment is simplistic and nearly tautological - It is not something difficult and subtle. It is a matter of recognizing *what* is being measured and scaled in each instance. "significant" means less likely to happen by chance, or, more deviant from the 'equality of means' or 'no-relation'. "stronger" means (a) bigger difference between groups, or (b) stronger association between continuous variables, when measured in the simple scale of differences. - So, 80 heads out of 100 is more unlikely, is a big effect compared to 50-50, and it is more significant than (say) 60/100. 80 is 30 percentage points (not zero), where 60 is 10; 80 is 8:2 or a 4:1 ratio (not 1:1), where 6:4 is 1.5:1. But, 4 out of 5 is still 80% so it is "stronger" than 60 out of 100, even though the latter may be rarer. For unlikelihood, or significance, '80 percent heads' (4 of 5) is easily recognized as easier to flip when it is 4 of 5, and not 80 of 100. [ ... ] -- Rich Ulrich, [EMAIL PROTECTED] http://www.pitt.edu/~wpilib/index.html - I need a new job, after March 31. Openings? - . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
