P
ParkerD
Guest
RebeccaJ,Well, it is at an undergrad level.
Which has nothing to do with your claim that they used an invalid statistical method. Still waiting for you to produce the evidence for this.
If their control selection is off, explain the statistical reasoning why, not just your “suspicions”.
The reason a scientific study establishes a control and does so as objectively as possible, is so that the bias of the researcher will not bias the results.
The reasoning why their method was incomplete and inadequate for saying they arrived at valid controls, is as follows:
- In the first place, they presented the wrong dates for Joel Barlow and the Columbiad. Barlow wrote the work (The Vision of Columbus, published in 1787) that he later used to make the Columbiad which he published in 1810. He died in 1812. The article says his work was written in 1825. This mistake shows the casual attitude these authors had in validating their controls.
- In the second place, the authors took “fifty novels of the same era” (American novels) and clustered those novels for similar word use, versus the cluster they came up with for the potential authors they had selected, and for the Book of Mormon. The cluster for the Book of Mormon itself was in a similar cluster to Barlow and Longfellow, but dissimilar to any of the other authors in the largest cluster set they ran. So what did they do? They came up with fewer words as a cluster set that would place their selected authors and the Book of Mormon and Barlow and Longfellow into a similar cluster set. (See the following link:
purl.stanford.edu/ir:rs276tc2764/index.html - By choosing to compare with “fifty novels of the same era”, they invalidated the idea of there being a control in their supposed “controls”, because they were now working with three different and distinct genres: the novel (they don’t describe which kind of novel, but most likely romance novels popular at the time); the epic poem; and a religious historical and personal essay narrative (The Book of Mormon).
- The authors should have taken a step back when they found that Longfellow was the closest match to being the writer of several chapters of the Book of Mormon using their statistical approach. (This is what a control is supposed to do–help the “scientist” take a step back from their arrival at conclusions when a finding invalidates their preconceived conclusion. During this step back, a rigorous scientist should expand their control set to see if they really had an adequate control set to begin with, and to see if they need to include other considerations that they had not previously addressed in establishing their control set.)
- Why no personal essay writers in their control set? Why no speeches in their control set, since the Book of Mormon presents a great number of “speeches” that are a genre in and of themselves (the closest other match being the personal essay)? This is why I suggested the use of Emerson and Thoreau, particularly Emerson.