4
$\begingroup$

EDIT: I was incorrectly looking to center my outcome variables. Only center predictors, and decide on group mean or grand mean centering by how you want to interpret your intercept.

I have 150 participants with 7 repeated-measures each that are equidistant(1 week apart) apart and set as time 0, 1, 2, 3, 4, 5, 6. I am having trouble with convergence of my models when adding parameters such as a random slope, I think it may be due to no centering. However, I do not know the manner in which to center these data properly.

Do I use the mean of all participants across all time points and center using that?

Or maybe the mean at each time point for all participants and center that time point with that mean?

Or do I take each participants mean across time and center all their points based on that?

Thanks for the help, I can find info about centering for traditional clustered data, but not for longitudinal data where participants are nested in time.

$\endgroup$

1 Answer 1

2
$\begingroup$

It is not clear from your description whether the 7 equidistant points are over time, i.e., whether you have longitudinal data.

If this is the case, then what sometimes helps is to change the scale of your time variable, e.g., from months to years.

$\endgroup$
3
  • 1
    $\begingroup$ Sorry, the measure is repeated over time and set as 0, 1, 2, 3, 4, 5, 6. $\endgroup$ Commented Feb 17, 2019 at 13:49
  • 2
    $\begingroup$ Then we need a bit more information like the output of the model even if it does not converge and a figure of the subject-specific longitudinal trajectories over time. $\endgroup$ Commented Feb 17, 2019 at 16:09
  • 1
    $\begingroup$ Sorry, I realized that I was looking to center my dependent variables, and you would not do that. The trouble of convergence came only when I added Time|Participant as a random effect, indicating a random slope. Therefore, I just conclude that my participants do not have different slopes, but only different intercepts. Thanks for the help! $\endgroup$ Commented Feb 21, 2019 at 14:51

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.