7090Re: [Synoptic-L] Some numerical results
- Dec 2, 2001At 10:43 AM 12/2/2001 +0000, Ron Price wrote:
> What surprises me about this whole discussion is that you have beenIn this regard, it occurs to me that the 95% level of assessing
>trying to make sense of a set of results which represents a small
>proportion of the published data, and apparently (unless I've missed
>something) without any attempt to check for overall consistency.
significance is far too generous. 95% implies that one in twenty
may be wrong, but in our data set we have 19 total sections (222
through 002) for a total of 171 different pair-wise comparisons
(19*18/2). If 1 in 20 is wrong (inherent in the 95% level), then
about 9 of the "significant" correlations are really just to due
to random chance.
This is why in such analyses textbooks recommend using P < 0.05/C,
which, in this case, works out to P < 0.0003. Thus, only those
correlations at that level should be considered significant. And
any attempts to build a theory should only be concerned with
Stephen C. Carlson mailto:scarlson@...
Synoptic Problem Home Page http://www.mindspring.com/~scarlson/synopt/
"Poetry speaks of aspirations, and songs chant the words." Shujing 2.35
Synoptic-L Homepage: http://www.bham.ac.uk/theology/synoptic-l
List Owner: Synoptic-L-Owner@...
- << Previous post in topic Next post in topic >>