Topic:
The same four proportional weighting factors work for each 00/01/10/11 when 0.25 is subtracted from each !!!
Replies:
506
Last Post:
Nov 20, 2012 9:21 PM
Although the expansion of our random control sets from 1 to 3 has the serendipitous effect of ameliorating at least the most egregious portion of our original "N-problem", I need to point out that this expansion is not only scientifically justifiable but scientifically obligatory.
On the non-random side, we started with the two original sets (63 S, 63 C) and:
i) createe two new sets (60 S, 60 C) by the well-known and scientifically meaningul operation of "reverse complementing" (63 S) to get (60 S) and then creating (60 S) from (60 C) the same way as we created (63 C from 63 S);
ii) created two new sets (119 S, 119 C) by forming (119 S) as the union of (63 S) and (60 S), and then creating (119 C) from (119 S) in the same way as (63 C) was created from (63 S) and (60 C) was created from (60 S).
But on the random side, we only started with the two original sets (R63 S, R63 C) parallel to (63 S, 63 C), i.e. we never constructed sets parallel to (119 S, 119 C) and (60 S, 60 C).
This is the deficiency that has now been corrected by the addition of:
iii) (R120, S R120 C) parallel to (119 S 119 C)
iv) (R58 S, R58 C) parallel to (60 S, 60 C).
Without the addition of these additional random control sets, both Jacques and Arthur will confirm that any scientific referee would question our design on the purely scientific grounds that we don't have "random apples vs non-random apples" for all three kinds of "non- random apples" that we've picked..
And, in addition to correcting a scientific design problem, the addition of the new sets serendipitously ameliorates at least the egregious portion of our "N-problem" caused by the comparisond of data involving (119 S, 119 C) to (R63 S, R63 C) that we executed because we hadn't bothered to construct (R120 S, R120 C). (And to a lesser extent, the comparison of (60 S, 60 C) to (R63 S, R63 C).
(Part of the problem here was my failure to fully automate earlier rather than later -prior to full automation, it would have been laborious to process data resulting from the additional sets (R120 S, R120C) and (R58 S, R58 C).)
