On 7 Nov 2002 09:42:55 -0800, [EMAIL PROTECTED] (rob) wrote: > I have a series of data coming from an instrument, (e.g. > spectrophotometer), let’s say 10 wavelenghts and 5 repetitions > of the measure. > [snip, much detail. Question: seems to be a difference between across-by-down, versus down-by-across ] > > > ex:data > 1 2 3 4 5 6 7 8 9 10 > 11 12 13 14 15 16 17 18 19 20 > 21 22 23 24 25 26 27 28 29 30 > 31 32 33 34 35 36 37 38 39 40 > 41 42 43 44 45 46 47 48 49 50 > > > choice1 > 21 22 23 24 25 26 27 28 29 30 > mean=255 > > > choice2 > 55 > 155 > 255 > 355 > 455 > mean=1275
Choice 1 of 255 is the sum of a bunch of means. Choice 2 of 1275 is the sum of a bunch of *Sums*. Multiplying by 5 fixes that discrepancy. On the other hand, you have several choices if you want describe the *variation* (Standard deviation) of certain quantities. You might have seen someone describe your present circumstance as a problematic one. If you look at those data as a 2-way ANOVA model, then you do have means (or sums) of values; rows; or columns; or cells. There is a "Total sum of squares" which would give you the standard deviation (SD) of individuals. "Mean squares", you should note, are estimates of deviations. You can subtract out the row means or column means, to get variations Within row or column; or subtract out both. -- Rich Ulrich, [EMAIL PROTECTED] http://www.pitt.edu/~wpilib/index.html . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
