Matthew L. Jockers follows up his 2008 study with a new study defending his findings.
Matthew L. Jockers, "Testing Authorship in the Personal Writings of Joseph Smith Using NSC Classification," Literary and Linguistic Computing 28, no. 3 (September 2013), 371–381
At the end of an experiment such as this, one would hope to have results that are far less messy, results that might lead to an obvious conclusion. The only conclusion comfortably reached here is that the decision to exclude Smith from prior work investigating the authorship of the Book of Mormon was entirely justified. Smith’s collected personal writings possess a great deal of stylistic variation—even among the works in his own hand—and the overall irregularity of the Smith corpus seems consistent with what one would expect from a set of co-authored documents. These results are also consistent with the conclusions of Mormon scholar Dean Jessee, who notes that ‘the impressions of Joseph Smith given [in the personal writings]... probably reflect the personality of the editor[s] more than they do Joseph’s’ (Smith and Jessee, 2002). Based on this research, and on Jessee’s historical work, I continue in the belief that the personal writings do not constitute a genuine sample of Smith’s linguistic style, and further, that the writings are likely to be exactly what Jessee says they are: documents that reflect the spirit of the man if not his style. Nothing in this current research moves me to change my opinion about the necessity of excluding Smith as an author-candidate in Book of Mormon authorship analysis, at least until such time as new, authenticated Smith documents are available. While I do find it tempting to do as suggested above and develop a Smith model based on the texts in his hand and those assigned to him in this experiment, I believe, given the variety of the NSC assignments, that it would be best to resist that temptation.