[R-sig-ME] following up reply regarding false positives

[John Kingston]

Dear Ben,

Thanks for your reply. I've responded to your suggestions below  
between *** ... ***


    The test you show below (although solve(Hessian,grad) is slightly
better) suggests that the first warning is indeed a false positive.

*** I ran the second test you suggest and reported the results in  
another post to this list with the subject heading "update to previous  
message about false positives". Here's the result again:

relgrad <- with(sfHSxRtCxRtR at optinfo$derivs, solve(Hessian, gradient))
max(abs(relgrad))

yielded a small value, too: 0.0003628809

Further tests following your suggestions to other who have posted to  
this list on this warning are also included in that message.

I've also run the allFit.R function on this model that you posted in  
yet another message. The outcome was OK for all optimizers. And this  
has been true for models of other data that have produced the same  
warning. ***

  I
would expect the second message to go away if you scale (centering is
probably a good idea too) your continuous predictors.  (The t-statistics
shouldn't change.)

*** I routinely center my predictors, and that was done in the model  
this message is about. Scaling is not an adjustment I'm familiar with.  
Should the continuous predictors be scaled down or up?

Current values of continuous predictors:

cStep:  -19 -15 -11  -7  -3  -1   1   3   7  11  15  19
cQRT:   -9 -5 -1  3  7  9

h1Context and h2Context are Helmert contrasts:

h1Context: 6 -3 -3
h2Context: 0 1 -1

***

Thanks again.
Best,
John

John Kingston
--------------------------------
Professor and Head
Linguistics Department
University of Massachusetts
150 Hicks Way, 226 South College
Amherst, MA 01003-9274
1-413-545-6837, fax -2792
jkingston at linguist.umass.edu
https://blogs.umass.edu/jkingstn/

John Kingston
--------------------------------
Professor and Head
Linguistics Department
University of Massachusetts
150 Hicks Way, 226 South College
Amherst, MA 01003-9274
1-413-545-6837, fax -2792
jkingston at linguist.umass.edu
https://blogs.umass.edu/jkingstn/