Re: Correlation of Sequence Order in GTh-Q1 Twin Sayings
- --- In email@example.com, "rj.godijn" <rj.godijn@...> wrote:
> a Mark/Q overlap means that Q theorists assume
> that the logion is found in both Q and Mark.
Of course that is consistent with the view that both Mark and Q derive
from Thomas. Maybe I should call this the Five Source Theory!
> I am not aware of any current Q econstruction thatI agree. I will revise the file I uploaded to include GTh 20 and GTh
> does not include the parable of the mustard seed...
> I think sticking to the schemas of Patterson and Mack
> means that you include GTh 20, because Patterson
> considers it a GTh/Q twin, but made a mistake in not
> including it as a Q parallel in his list (the same
> goes for GTh 94).
94. The revised Pearson correlation for Q1a has p <= 0.00444, and for
Q1b p <= 0.148. This would leave the regression for Q1a significant at
the 99% confidence interval, but that for Q1b not significant at the
95% confidence interval.
Unless, that is, a new series (Q1c) begins following GTh 96.
This would produce the following sequence:
GTh26 = QC1.01 = QS12 = Q1a
GTh34 = QC1.01 = QS11 = Q1a
GTh45 = QC1.01 = QS13 = Q1a
GTh54 = QC1.01 = QS8 = Q1a
GTh68 = QC1.01 = QS8 = Q1a
GTh69 = QC1.01 = QS8 = Q1a
GTh73 = QC1.02 = QS20 = Q1a
GTh86 = QC1.02 = QS19 = Q1a
GTh92 = QC1.03 = QS27 = Q1a
GTh94 = QC1.03 = QS27 = Q1a
GTh21 = QC1.03 = QS27 = Q1b
GTh33 = QC1.04 = QS35 = Q1b
GTh51 = QC1.04 = QS35 = Q1b
GTh61 = QC1.04 = QS35 = Q1b
GTh76 = QC1.05 = QS40 = Q1b
GTh96 = QC1.06 = QS46 = Q1b
GTh20 = QC1.06 = QS46 = Q1c
GTh55 = QC1.07 = QS52 = Q1c
GTh101 = QC1.07 = QS52 = Q1c
Here for each set - Q1a, Q1b and Q1c - an increase in GTh Saying Number
is associated with increase or sameness of Q1 Cluster Number.
Obviously there is a question of forcing linear sequences by starting
over when it is convenient to do so. This involves starting over twice
out of 18 opportunities. If this can be described by a binomial
distribution, then P <= 0.001 (0-2 successes, 18 trials,
TrialP(success) = 0.5. However this I think this underestimates P, as
for most data pairs (GTh, QC) the probability is greater than 0.5 that
another data pair has greater than or same Q Cluster Number. If the
effective TrialP(success) >= 0.31, then P <= 0.05. I will have to see
if I can find or simulate
a statistical method for this.
And of course I have to point out that the boundary between Q1b and
Q1c is another catch-saying, QS46!