[Rsiena-help] What does NA for standard errors mean?

[Tom Snijders]

Dear Tanmay,

For this type of question, the yahoo discussion list for Siena/Stocnet
http://groups.yahoo.com/groups/stocnet/
Is more appropriate (the yahoo list is meant for discussions between users, this list is for discussions between programmers).

But let me answer this here. The NA standard errors are given in case of parameter collinearity. In this case they mean that you have included a set of effects that together are linearly dependent: some linear combination of them has infinite standard error.
I suggest that you retain only one of the GWESP effects, and test the other three by a score-type test (manual section 8.2), and include them only (one by one in successive model fits) if from the score test it can be concluded that they are significant.

Best wishes,
Tom

================================================================
Tom A.B. Snijders
Professor of Statistics and Methodology
Dept. of Sociology, University of Groningen
Emeritus Fellow, Nuffield College, University of Oxford
Associate Member, Dept. of Statistics, University of Oxford
http://www.stats.ox.ac.uk/~snijders/

From: rsiena-help-bounces at lists.r-forge.r-project.org [mailto:rsiena-help-bounces at lists.r-forge.r-project.org] On Behalf Of tanmay sinha
Sent: 15 April 2015 16:45
To: rsiena-help at lists.r-forge.r-project.org
Subject: [Rsiena-help] What does NA for standard errors mean?

Hi all, I am a new user to RSiena. After running the model that includes certain structural configurations, I find that some of my standard errors are NA.
I include a portion of the output, with the 3 columns representing estimates, standard errors and t-ratios.

   8. eval GWESP I -> K -> J (69)               2.1135  (     NA   )   -0.7091
   9. eval GWESP I <- K <- J (69)               0.1239  (     NA   )   -0.1293
  10. eval GWESP I <- K -> J (69)              -0.4884  (     NA   )   -0.0660
  11. eval GWESP I -> K <- J (69)              -0.4797  (     NA   )   -0.1918


What do the NA mean?? How do I interpret this?

I would greatly appreciate your help.

Regards
Tanmay Sinha
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.r-forge.r-project.org/pipermail/rsiena-help/attachments/20150416/64a0ecb8/attachment.html>