From: Karen Pitts (karen_pitts@maca.sarnoff.com)
Date: Thu Oct 26 1995 - 23:05:47 EDT
Statistics on LEGW in the Gospels 10/27/95 10:17 AM
Thanks to Stephen Carlson for the data on the forms of LEGW in the Gospels
(it's dangerous to give a statistician data!).
The correct way to look at this is to scale the occurances by the total usage
of all forms of LEGW in each Gospel. Or we can crank the numbers through a
chi-squared test of independence and look at the deviations. Both are
presented below. Note that the Gospels are alphabetized, not in NT order.
I'm only including the table that has all the Gospels. The big differences
are that Luke uses EIPEN much more than the others (and Mark uses it less),
Mark uses ELEGEN more, Luke uses ELEGON less (and John more), Luke uses LEGEI
much less (and John more), Matthew uses LEGONTES more (and John less), Luke
uses LEGOUSIN much more (and Mark less), and Matthew uses LEGWN more (and John
less). I conclude from this that Luke is the most different, so I excluded
Luke and compared Matthew to Mark and John to Mark. I won't reproduce the
complete tables, but the major conclusion is that in Matthew vs. Mark, Matthew
prefers EIPEN, LEGONTES, LEGOUSIN while Mark prefers ELEGEN, ELEGON, and
LEGEI. For John vs. Mark, John prefers EIPEN and LEGOUSIN while Mark prefers
ELEGEN and LEGWN. All the usage differences are statistically significant.
In assessing differences I examined the big contributors to the chi-square.
If anyone wants more gory detail, let me know.
So I think that my observation is supported by the statistics. Mark uses
imperfect more than the others and uses aorist less.
Crosstabs
LEGW form Gospel
Count John Luke Mark Matt Row Totals
Col %
Deviation
EIPAN 26 28 9 16 79
5.80 4.4 3.2 3.7
6.3 0.1 -3.5 -2.9
EIPEIN 0 7 2 2 11
0.0 1.1 0.7 0.5
-2.7 3.1 0.3 -0.6
EIPEN 115 231 59 119 524
25.7 36.4 20.9 27.7
-15.8 45.9 -23.6 -6.5
ELEGEN 13 20 31 3 67
2.9 3.2 11.0 0.7
-3.7 -3.7 20.4 -13.1
ELEGON 35 4 19 8 66
7.8 0.6 6.7 1.9
18.5 -19.3 8.6 -7.8
LEGEI 123 14 62 54 253
27.5 2.2 21.9 12.6
59.8 -75.4 22.1 -6.6
LEGEIN 1 12 8 5 26
0.2 1.9 2.8 1.2
-5.5 2.8 3.9 -1.2
LEGONTES 10 36 15 47 108
2.2 5.7 5.3 10.9
-17.0 -2.1 -2.0 21.1
LEGOUSA 2 4 1 8 15
0.5 0.6 0.4 1.9
-1.7 -1.3 -1.4 4.4
LEGOUSIN 115 231 59 119 524
25.7 36.4 20.9 27.7
-15.8 45.9 -23.6 -6.5
LEGWN 8 47 18 49 122
1.8 7.4 6.4 11.4
-22.4 3.9 -1.2 19.8
Totals 448 634 283 430 1795
Tests
Source DF -LogLikelihood RSquare (U)
Model 30 190.5472 0.0570
Error 1755 3150.0799
C Total 1785 3340.6271
Total Count 1795
Test ChiSquare Prob>ChiSq
Likelihood Ratio 381.094 0.0000
Pearson 357.663 0.0000
Karen Pitts
Hopewell Presbyterian Church, Hopewell, NJ, teacher of NT Greek
David Sarnoff Research Center, Princeton, NJ, statistician
kpitts@sarnoff.com
This archive was generated by hypermail 2.1.4 : Sat Apr 20 2002 - 15:37:31 EDT