[R-sig-ME] upcoming release of lme4 1.1-7

Ben Bolker bbolker at gmail.com
Mon Jun 23 15:27:17 CEST 2014

   We are on the verge of submitting a new release of lme4 to CRAN.
Probably the most interesting aspect of this release from the end-user
perspective is that we have improved the convergence tests (following
much discussion on this list) so that we no longer expect
false-positive max|grad| warnings even on large problems.  Other than
that, this is largely a bug-fix release.  (We have added some
trouble-shooting information to the ?lmerControl help page, and hope
to expand this section in future releases.)

  We have not changed any of the optimization behaviour, but we are
strongly considering moving to using the nloptr implementation of
BOBYQA as the default optimizer in future versions (lme4 now Imports
the nloptr package).

  We don't expect any compatibility problems with downstream packages,
but as always it would be a good idea for downstream package
maintainers to test with the new version (from Github).

    Ben Bolker, on behalf of the lme4 maintainers



         • the ‘nloptr’ package is now imported; a wrapper function
           (‘nloptwrap’) is provided so that
           ‘lmerControl(optimizer="nloptwrap")’ is all that's necessary
           to use ‘nloptr’ optimizers in the nonlinear optimization
           stage (the default algorithm is NLopt's implementation of
           BOBYQA: see ‘?nloptwrap’ for examples)

         • preliminary implementation of checks for scaling of model
           matrix columns (see ‘check.scaleX’ in ‘?lmerControl’)

         • ‘beta’ is now allowed as a synonym for ‘fixef’ when
           specifying starting parameters (Github #194)


         • the use of ‘deviance’ to return the REML criterion is now
           deprecated; users should use ‘REMLcrit()’ instead (Github

         • changed the default value of ‘check.nobs.vs.rankZ’ to
           ‘"ignore"’ (Github #214)


         • change gradient testing from absolute to relative

         • fix ‘confint(.,method="boot")’ to allow/work properly with
           ‘boot.type’ values other than ‘"perc"’ (reported by Alan

         • allow ‘plot()’ to work when data are specified in a
           different environment (reported by Dieter Menne)

         • ‘predict’ and ‘simulate’ work for matrix-valued predictors
           (Github #201)

         • other ‘simulate’ bugs (Github #212)

         • ‘predict’ no longer warns spuriously when original response
           was a factor (Github #205)

         • fix memory access issues (Github #200)

More information about the R-sig-mixed-models mailing list