you can just explain that you're doing better than listwise deletiion.
you can't really provide evidence against nonignorability from your data
set (by definition of nonignorability). but you know that its a better
assumption than listwise if lthe 2 give different answers. so a very
brief discussion; you do not need to repeat things in prior literature.
just give the reader the results and refer them to the original paper.
Gary
On Sun, 11 May 2003, Beau Kilmer wrote:
I'm using AMELIA and wonder how much of a
discussion I should include
about "missingness assumptions." Should I just refer to King et al.
(2001), or should I include a few paragraphs giving the reader some
background?
For what it's worth, the audience for my paper is economists and
criminologists.
Thanks.