Interpreting graph results:

Reply

Jo Decaesteker | Forum Activity | Posted: Fri, May 24 2013 6:14 AM

Hi,

I'm hoping to share insights, but also as some questions here.

I'm in the process to identify keywords for the gospel by Mark as a whole. (So, not for a particular passage/chapter but for the whole gospel of Mark!) I'm now looking at one specific example, "λεγω" (as a root). (This may be far from a keyword, but let's just take this example with λεγω, so I can ask my questions...)

I know that mathematics and sheer numbers are not the only answer and I should not reduce it to that, but an important factor (one of several criteria!) in determining whether a word/root could be a keyword for a Biblebook as a whole(!), is its distribution within the book: is it spread evenly over all/most of the chapters, or is it more clustered in one or more passages? And what is its relation to each chapter (ratio) and/or the whole book?

The graph results are a great help for this.

First of all, the "Result Map" gets me a good view on how the word/root is spread all over the entire book (you can zoom in and out by scrolling!): (I'm looking at Mark, which is the second gospel, so logically, the second column is colored).

Second, I can analyze per chapter/verse... with the "Bar Chart" view:

Can anyone help me with interpreting the several possible views within the Bar Chart? When is each method of viewing most suitable? I'm especially looking at the views with a ratio: "Number of... / Number of ....."

Cheers!

Jo D.

Reply

Mark Hoffman | Forum Activity | Replied: Fri, May 24 2013 10:05 PM

Someone can certainly correct me on this, especially the terms I will use, but I think that there are three ratios that are most significant. As an example, I'm using occurrences of "truth" in the NRSV NT.

1. Number of Hits in Book / Number of Words in Book: I'll use the word "density" or "saturation" for this ratio. It gives a good picture of how 'saturated' a book is with that particular word. Regarding "truth," 3John is most saturated followed by 2John then 1John and much further down the line is John. (If you want a visual comparison, think of every hit as a light. With this ratio, every book is like a bulb, and its size is relative to the size of the book. How often the hit occurs is indicates how bright the bulb is, so in this case, 3John may be a small bulb but it is an intense light. John is a big bulb but not nearly as bright.)
2. Number of Hits in Book / Number of Words in Version: I call this a measure of "overall frequency," since now the hits in the book are measured against all the words in all the books. Now John is by far the most common followed by 1John and somewhat further down, 2John and 3John. Note that this is basically the same as Number of Hits in Book / Number of Hits in Version. The ratios will be different numbers, but the relative frequencies are identical. (Using the visual analogy again, this time every hit/light in a book is placed into the same size bulb. Now you can see why John will be so much brighter than 3John.)
3. The only other ratio I sometimes use is one of the Number of Hits in Chapter. If you enable the "Show zero items" it will give you the sense of the frequency of the hit within the book.

Reply

Jo Decaesteker | Forum Activity | Replied: Sun, May 26 2013 11:43 PM

Thanks, Dave, Mark, M.J., good insights.

My goal is to come a little closer to determining potential keywords in a book as I said.

M.J.: good point about the chapters: these are constructs of later date: the authors didn't have those in mind when writing their letter(s)/book(s) indeed.

Could it be interesting to have a customizable way to search for a ratio, like "Number of Hits / X words in Book", where X can be set between certain minimum/maxium like 100-1000...? This might give a possibly more precise look at the distribution (especially for a ratio with a smaller "X", like 100)...