fixing the value of some parameters in an lme or lmer model

Dear Asher,

You start from a model with this equation $y = \beta_0 + \beta_1 week
+ b_{i0} + b_{i1} week$
The fit of the baseline of each patient is $baseline_i = \beta_0 +
b_{i0}$ but you want $baseline_i = b_{i0}$ hence $\beta_0 = 0$
Forcing a parameter to be 1 can be done with offset(): $y = 0 +
\beta_1 week + offset(baseline_i) + b_{i1} week$

The lmer formula becomes y ~ offset(baseline) + week + (0 + week |
subject), assuming that the baseline has week = 0 and week is
continuous

Best regards,

ir. Thierry Onkelinx
Instituut voor natuur- en bosonderzoek / Research Institute for Nature
and Forest
team Biometrie & Kwaliteitszorg / team Biometrics & Quality Assurance
Kliniekstraat 25
1070 Anderlecht
Belgium

To call in the statistician after the experiment is done may be no
more than asking him to perform a post-mortem examination: he may be
able to say what the experiment died of. ~ Sir Ronald Aylmer Fisher
The plural of anecdote is not data. ~ Roger Brinner
The combination of some data and an aching desire for an answer does
not ensure that a reasonable answer can be extracted from a given body
of data. ~ John Tukey


2017-03-29 10:39 GMT+02:00 Asher Strauss <asher.strauss at gmail.com>:

Dear list,

I am trying to fit a two level hierarchical model using the lme/lmer
function.
I have two levels in my data: 3 time points (codded: week 0, week 4, week
8) nested within subjects (ID). I would like to set the intercept of each
patient to the value of the outcome at week 0 (the first observation), or
in other words force the model fit for each subject to pass through the
subject's baseline score on the outcome variable.

Here are the formulas (in informal notation):

Level 1: (time)
outcome=b0+b1*week+error

Level 2: (subjects)
b0=*0*+*1**"outcome-at-time-0"
b1="grand-slope"+error

I underlined and made bold the values I would like to force.

Can this be done in nlme or lme4 packages?

Thank you very much!!!

Asher Strauss

        [[alternative HTML version deleted]]

_______________________________________________
R-sig-mixed-models at r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-sig-mixed-models