I understand that problem, but I think such a study could have been published in an academic journal if the author had access to the original text. In this case, however, that is impossible.
________________________________________________________________________________________________
John Hilton and the Berkeley Group
Methods
A more sophisticated approach was taken by John Hilton and non-LDS colleagues at Berkeley.
[6] The "Berkeley Group's" method relied on non-contextual word
patterns, rather than just individual words. This more conservative method was designed from the ground up, and required works of at least 5,000 words.
The Berkeley Group first used a variety of control tests with non-disputed authors (e.g. works by Mark Twain, and translated works from German) in an effort to:
- demonstrate the persistence of wordprints despite an author's effort to write as a different 'character'
- demonstrate that wordprints were not obliterated by translation (e.g. two different authors rendered by the same translator would still have different wordprints).
The Berkeley Group's methods have since passed peer review, and were used to identify previously unknown writings written by Thomas Hobbes.
[7]
The Berkeley Group compared Book of Mormon texts written by Nephi and Alma with themselves, with each other, and with work by Joseph, Oliver, and Solomon Spaulding. Each comparison is assessed based upon the number of "rejections" provided by the model. The greater the number of rejections, the greater the chance that the two texts were
not written by the same author. Tests with non-disputed texts showed that two texts by the same author never scored more than 6 rejections; thus, one cannot be certain if scores between 1–6 were written by the same or different authors. Scores of 0 rejections makes it statistically likely the two texts were written by the same author.
However, seven or more rejections indicates that the texts were written by a different author with a high degree of probability:
[8]
# of Rejections Certainty of being
different authors
7 99.5%
8 99.9%
9 99.99%
10 99.997%
Results
The results are striking:
[9]
Recall that any test over 6 indicates different authorship; 1–6 or less is indeterminate; 0 is same author. Each x represents one test.
(This is a table on the website linked to below so I can't get it to look right here. However, it should serve to show that there are few 'rejections' among the authors compared to themselves and many many rejections among the authors when compared to different authors.)
Total Number of
Tests Performed
0 1 2 3 4 5 6
7 8 9 10 11 12 13 14 15
Nephi vs. Nephi 3 ---- ---- x ---- x x ---- ---- ---- ---- ---- ---- ---- ---- ---- ----
Alma vs. Alma 3 ---- x x x ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ----
Smith vs. Smith 3 x ---- xx ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ----
Cowdery vs. Cowdery 1 ---- x ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ----
Spaulding vs. Spaulding 1 ---- ---- x ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- ----
Nephi vs. Alma 9 ---- ---- x ---- ----xx xx x x x x ---- ---- ---- ---- ----
Smith vs. Nephi 6 ---- --- --- --- x ---- ---- ---- xx --- x x x ---- ---- ----
Smith vs. Alma 6 ---- ---- ---- xx x x ---- xx ---- ---- ---- ---- ---- ---- ---- ----
Cowdery vs. Nephi 6 ---- ---- ---- ---- ---- ---- x x --- --- ---- xx ---- x x ----
Cowdery vs. Alma 6 --- --- ---- ---- ---- ---- ---- xxxx x x ---- ---- ---- ---- ---- ----
Spaulding vs. Nephi 6 ---- ---- ---- ---- ---- ---- ---- ---- ---- ---- x x x ---- x xx
Spaulding vs. Alma 6 ---- ---- ---- --- --- --- xxx ---- xx ---- ---- ---- x ---- ---- -
Furthermore, each "rejection" is statistically independent—this means that the chance of two different authors being the product the same person can be determined by
multiplying the chance of each individual failure.
[10]
Thus the chance of Nephi and Alma being the same author is found by:
chance of 7 rejections x 8 rejections x 9 rejections x 10 rejections
= 0.005 x 0.001 x 0.0001 x 0.00003
=
0.000000000000015
=1.5 x 10-14
This is a roughly 1 in 15
trillion chance of Nephi and Alma having the same author. Hilton rightly terms this "statistical overkill".
Authors Cumulative chance of being the same author
Nephi and Alma 1.5 x 10^-14
Joseph Smith and Alma 2.5 x 10^-5
Joseph Smith and Nephi less than 2.7 x 10^-20
Oliver Cowdery and Alma 6.25 x 10^-17
Oliver Cowdery and Nephi less than 8.1 x 10^-19
Spaulding and Alma less than 3 x 10^-11
Spaulding and Nephi less than 7.29 x 10^-28
___________________________________________________________________________________________
This is from a longer article at
http://en.fairmormon.org/Book_of_Mormon/Wordprint_studies
I find it interesting that the methods for this particular study have since been peer reviewed. Also that the statement that going from one language to another does not negate wordprint evidence. Finally, this study was performed by non-LDS scholars. That should make it worth more to those who complain about Mormon bias.