There is a historical debate about the role of Mark Twain in the Civil War. During 1861, a set of ten letters signed eOuintus Curtius Snodgrass had appeared in the New Orleans Daily Crescent. The letters purported to Twains military achievements as a member of the Louisiana milita. While historians generally agree that the accounts referred to actually did happen, there seems to be no record of anyone named Quintus Curtius Snodgrass. (Larsen & Stroup 1976) The hypothesis is that Twain and Snodgrass could have been the same person.
In order to investigate this matter, one of the best ways is to examine both of their writings. Firstly, we need to carefully choose their sample writings, and examine word length frequencies. This approach would possibly give a view if Twain and Snodgrass are the same because [a] given author will use roughly the same proportion of, say, three-letter words in something he writes this year as he did in whatever he wrote last year cBut, the proportion of three-letter words that Author A consistently uses will very likely be different than the proportion of three-letter words that Author B uses. Theoretically, then, by constructing a word-length frequency count for essays known to be written by Mark Twain and comparing that to a similar count made for the Snodgrass letters, it should be possible to assess the likelihood of the two authors being one and the same. (Larsen & Stroup 1976) The great difference of the word frequency distributions for two sets of writings woul!.
d be enough evidence that different people wrote the two sets. On the other hand, it does not prove that the same person writes those two writings even if the distributions are quite similar because it is very possible for two different authors to write in similar styles. Next approach is to examine Chi-square that a standard statistical procedure for measuring the difference between two frequency distributions.
