1
$\begingroup$

I'm quite new to the concept of correlated data still, and I was have a few questions about unbalanced longitudinal data.

Let's say we are using a linear mixed model:

1) what is the impact of not adjusting the unbalanced nature of time?

2) what would we be assuming if we did not adjust?

3) what would be the assumptions if we did adjust for the unbalanced nature?

3) is it critical to adjust for unbalanced time?

Thank you very much!

EDIT for my situation Y is a repeated measure, but the time at which it's measured is unbalanced for the participants

Improve this question
$\endgroup$
2
  • $\begingroup$ Linear Mixed Model: $Y=X\beta+Z\gamma+\epsilon, \gamma \sim N(0,G), \epsilon \sim N (0,R)$. What part(s) is your (un)balanced nature involved? $\endgroup$
    – user158565
    Commented Dec 11, 2018 at 0:23
  • $\begingroup$ @user158565 for my situation Y is a repeated measure, but the time at which it's measured is unbalanced for the participants $\endgroup$
    – j681
    Commented Dec 11, 2018 at 0:32

1 Answer 1

Reset to default
2
$\begingroup$

Linear mixed models can work with unbalanced data, i.e., there is no problem if subjects are measured at completely different time points. You do not need to group time points together to create discrete follow-up times. In the specification of the model you may need to account for potentially nonlinear profiles over time, using splines or polynomials.

Improve this answer
$\endgroup$
2
  • $\begingroup$ Is there a statistical proof for this? $\endgroup$
    – user430997
    Commented Oct 3 at 4:32
  • $\begingroup$ This stems from how mixed models are defined and estimated. You work in the long format with the measurements of different subjects stacked underneath each other. This does not require subjects to have the same number of measurements or be measured at the same time points. $\endgroup$
    – Dimitris Rizopoulos
    Commented Oct 3 at 8:25

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.