[R-sig-ME] bivariate growth models using lme()

Simon Harmel @|m@h@rme| @end|ng |rom gm@||@com
Tue Jan 19 03:29:22 CET 2021


Hello all,

In chapter 8 of this book (
https://www.guilford.com/books/Growth-Modeling/Grimm-Ram-Estabrook/9781462526062),
p. 172, script 8.2, there is a demonstration of a bivariate growth model
fit with `nlme()` function (see `m1` below).

I was wondering if the same model could be fit using the `lme()` function?
I have tried `lme()` (see `m2` below), but the output doesn't match. --
Thanks, Simon

multivariate <- read.csv('
https://raw.githubusercontent.com/hkil/m/master/bv.csv')


lmeCtlList <- lmeControl(maxIter = 200, msMaxIter = 200, niterEM = 50,
           msMaxEval = 400)

m1 <- nlme(var~d_math*(b_1i+b_2i*(grade-2))+
         d_hyp *(h_1i+h_2i*(grade-2)),
       data      = multivariate,
       fixed     = b_1i+b_2i+h_1i+h_2i~1,
       random    = b_1i+b_2i+h_1i+h_2i~1,
       group     = ~id,
       start     = c(35, 4, 1, -1),
       weights   = varIdent(c(hyp=.3), form = ~1|grp),
       na.action = na.omit,
       control=lmeCtlList)


m2 <- lme(var ~ 0 + d_math + d_hyp + d_math:grade + d_hyp:grade,
    random = ~ 0 + d_math + d_hyp + d_math:grade + d_hyp:grade | id, data =
multivariate,
    na.action = na.omit, weights  = varIdent(c(hyp=.3), form = ~1|grp),
    control = lmeCtlList)

# `m1` output:
coef(summary(m1))
           Value  Std.Error   DF   t-value       p-value
b_1i 35.25857425 0.35511071 3456 99.288962  0.000000e+00
b_2i  4.34307216 0.08737929 3456 49.703680  0.000000e+00
h_1i  1.90351659 0.05815329 3456 32.732739 6.143241e-205
h_2i -0.05663496 0.01422441 3456 -3.981533  6.988077e-05

  # `m2` output:
  coef(summary(m2))
                   Value  Std.Error   DF   t-value       p-value
d_math       26.57270335 0.47201444 3456 56.296378  0.000000e+00
d_hyp         2.01699310 0.07867055 3456 25.638478 7.225043e-133
d_math:grade  4.34308917 0.08740932 3456 49.686797  0.000000e+00
d_hyp:grade  -0.05670937 0.01414676 3456 -4.008647  6.235700e-05

	[[alternative HTML version deleted]]



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