# [NMusers] RE: Standard errors of estimates for strictly positive parameters

From: Bob Leary <Bob.Leary_at_certara.com>
Date: Thu, 12 Feb 2015 14:04:02 +0000

Dear Aziz -
The approximate likelihood methods in NONMEM such as FO, FOCE,and LAPLACE =
optimize an objective function than is parameterized internally
by the Cholesky factor L of Omega, regardless of whether the matrix is diag=
onal (the EM -based methods do something considerably different and work di=
rectly with Omega rather than
the Cholesky factor.)

Thus for the approximate likelihood methods, the SE's computed internally b=
y \$COV from the Hessian or Sandwich or Fisher score methods
are first computed with respect to these Cholesky parameters , and then the=
corresponding SE's of the full Omega=LL' are computed by a 'propagation =
of errors' approach
which skews the results, particularly if the SE's are large. Thus in a sen=
se regarding your dilemma about whether Model 1 or Model 2 is better with r=
espect to applicability of \$COV results, one answer is that both are fundam=
entally distorted by the propagation of errors method with respect to the =
Omega elements.

But regarding your fundamental question 'can we trust the output of \$COV '=
- all of this makes very little difference. Standard errors computed by \$C=
OV are inherently dubious - the applicability of the usual asymptotic argum=
ents is very questionable for the types/sizes of data sets we often deal wi=
th.
As Lewis Sheiner used to say of these results, 'they are not worth the elec=
trons used to compute them'. They are the best we can do for the level o=
f computational investment put into them -
If you want something better, try a bootstrap or profiling method.

________________________________________ NOTICE: The information co=
ntained in this electronic mail message is intended only for the personal a=
nd confidential use of the designated recipient(s) named above. This m=
essage may be an attorney-client communication, may be protected by th=
e work product doctrine, and may be subject to a protective order. As such,=
this message is privileged and confidential. If the reader of this me=
ssage is not the intended recipient or an agent responsible for delivering =
it to the intended recipient, you are hereby notified that you have re=
ceived this message in error and that any review, dissemination, distr=
ibution, or copying of this message is strictly prohibited. If you have rec=
eived this communication in error, please notify us immediately by tel=
ephone and e-mail and destroy any and all copies of this message in yo=
ur possession (whether hard copies or electronically stored copies). Thank =
you. buSp9xeMeKEbrUze

Received on Thu Feb 12 2015 - 09:04:02 EST

This archive was generated by hypermail 2.3.0 : Fri Sep 27 2019 - 16:43:03 EDT