From: KAREN PITTS (kpitts@sarnoff.com)
Date: Thu Jun 17 1999 - 22:52:24 EDT
<x-rich>The correct statistical test is a Chi-Squared test of independence,
which basically tests whether the proprotions are all the same for the
two columns. I did the test for the entire table, which is highly
statistically significant (error << e-27), the table without SATANAS,
which is highly statistically significant (error << e-15), and for the
table without SATANAS and PETROS, which is not statistically
significant (error = .12).
I used the frequencies, which I've pasted below. The proportions are
from Steven Cox's original request.
Now, whether the statistical significance has any meaning is a question
I can't answer.
Karen
<fontfamily><param>Courier</param><bigger> with art. without
totals
Abraham 12 60 72
Moses 13 66 79
David 5 53 58
Elijah 0 29 29
John 4 30 34
Petros 52 40 92
Satanas 33 4 37
totals 119 282 401
</bigger></fontfamily>Word total %w.art. %w/o art.
Abraham 72 16.4 83.6
Moses 79 16.46 83.54
David 58 8.6 91.40
Elijah 29 0 100
John 34 11.8 88.2
Petros 92 56.5 43.5
Satanas 37 86.49 13.51
Karen Pitts
statistician at Sarnoff Corporation, Princeton NJ
teacher of NT Greek, Hopewell Pres. Church, Hopewell, NJ
</x-rich>
This archive was generated by hypermail 2.1.4 : Sat Apr 20 2002 - 15:40:31 EDT