Λύχνις Δαν Posted March 16, 2019 Share Posted March 16, 2019 (edited) Hi ya, Acc 12.3.3rc3 Win 10 - I have not tested this on Mac but it would surprise me if this is not generic. Open the NA28 Greek Open the Table analytics 1. Search for * [RANGE acts ] -> should show 18472 hits - the total number of words in the searched range Table will show hits/1000 -> 1000. Correct. 2. Search for μεν [RANGE acts 1-15:35] -> should show 23 Table will show hits/1000 -> 1.25 Incorrect 3. Search for * [RANGE acts 1-15:35] -> should show 10063 hits Table will show hits/1000 -> 544.77 Incorrect The problem is this: the denominator in the hit/1000 for all three of these queries is the same, the number of words in the whole of Acts. The correct answer for query 2 is 23 / 10063 * 1000 = 2.286 The correct answer for query 3 is 1000, just as it is for query 1. Thx D Edited March 16, 2019 by דָנִיאֶל Link to comment Share on other sites More sharing options...
Helen Brown Posted March 16, 2019 Share Posted March 16, 2019 I do not think that the Table has ever given the frequency based on the search range. It is designed to calculate the frequency based on entire book or chapter. Link to comment Share on other sites More sharing options...
Λύχνις Δαν Posted March 16, 2019 Author Share Posted March 16, 2019 (edited) The problem with that is that the numerator and denominator are then inconsistent with one another for searches specifying a different range, as here. That renders the statistics incorrect though it's not clear that they are from the chart. I only actually figured this out because I was comparing the numbers with others in an article and had cause to recalculate them myself from the raw counts. At that point it became clear something was amiss. Incidentally, the numbers in the article mostly came from Accordance 8 (hence my other question about getting a search done in Acc 8 the other day). Now those hits/1000 numbers are noticeably closer to mine than to Acc's current numbers. Not having access to an Acc 8 env I cannot determine the cause. I can obviously workaround this now I know it happens. I can just extract raw counts and calculate my own ratios, but I figured I would mention the bug. Thx D Edited March 16, 2019 by דָנִיאֶל Link to comment Share on other sites More sharing options...
Recommended Posts
Please sign in to comment
You will be able to leave a comment after signing in
Sign In Now