GalacticCactus Forum

Forums => English & Linguistics => Topic started by: Jonathon on September 23, 2009, 11:27:12 AM

Title: Statistics editing help
Post by: Jonathon on September 23, 2009, 11:27:12 AM

I'm editing a book that makes copious use of statistical data, but an awful lot of that data looks fishy. Unfortunately, I have no training in statistics, so I don't know how to fix the problems. For instance, I've got a table that looks something like this:

Code: [Select]

R²        .494     .634     .390     .285     .314     .379
?²         475      279      288     .230      239      299
df         178      176       90     .176      145      162
DLK       .878     .890     .899     .953     .903     .943

This same issue—where the figures in two rows lack decimals except in one column—is repeated on several tables. I just don't know whether they should all be decimals or not.

Title: Statistics editing help
Post by: The Genuine on September 23, 2009, 11:35:47 AM

Can you not ask the author?

Title: Statistics editing help
Post by: Jonathon on September 23, 2009, 11:39:23 AM

I can, but it will take a while to receive a response. I thought polling the audience might be quicker.

Title: Statistics editing help
Post by: fugu13 on September 24, 2009, 09:59:27 AM

That's strange, the R squared and Chi squared rows should both have decimal points everywhere, but the degrees of freedom (df) row should have decimal points nowhere (degrees of freedom are always integers).

DLK isn't a statistical term I (or Google) have any familiarity with. I suspect it is specific to the model they are talking about, so the decimal points are almost certainly appropriate.

Title: Statistics editing help
Post by: Jonathon on September 24, 2009, 10:10:41 AM

Thanks, fugu!

The book is pretty sloppy in a lot of areas, so, unfortunately, I'm not surprised that they messed up with the decimals. It doesn't exactly inspire a lot of confidence in the rest of the book, of course.

Title: Statistics editing help
Post by: Jonathon on October 07, 2009, 12:07:47 PM

Well, the author wrote back and said that neither the chi squared nor df figures should have decimals. So now I'm really confused, and I suspect that these authors don't really know what they're doing.