turn on suggestions

Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

Find a Community

- Home
- /
- Analytics
- /
- Stat Procs
- /
- Quadratic time-varying predictor in MIXED model?

Topic Options

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

02-10-2011 06:40 PM

Hello,

I'm working on analysing longitudinal data with various starting ages and age between time-points as well as some missing data, so I'm employing a mixed time-varying model, looking to assess if my dependent (number of words spoken) change is associated with changes in a time-varying predictor (IQ) (i.e. IQ is assessed at each time point). I had assumed that there was no systematic change in IQ with time (there really shouldn't be!), but looking at the spaghetti plot, there is a very clear quadratic curve, such that IQ drops at a certain age, and then goes back up- regardless of what this actually means (it could just be a coincidence), I'm concerned that it will be messing up my model. As is, I was using starting-age, age-from-time1, gmc-centered starting-iq, and PMC-centered change -in-iq-from-time1 as my fixed factors, as well as a quadratic factor of age-from-time1 (and Iq's interactions with time). Can I simply include my IQ variables interacting with the quadratic time function, or is there something more complicated that I need to do?

Thanks!

I'm working on analysing longitudinal data with various starting ages and age between time-points as well as some missing data, so I'm employing a mixed time-varying model, looking to assess if my dependent (number of words spoken) change is associated with changes in a time-varying predictor (IQ) (i.e. IQ is assessed at each time point). I had assumed that there was no systematic change in IQ with time (there really shouldn't be!), but looking at the spaghetti plot, there is a very clear quadratic curve, such that IQ drops at a certain age, and then goes back up- regardless of what this actually means (it could just be a coincidence), I'm concerned that it will be messing up my model. As is, I was using starting-age, age-from-time1, gmc-centered starting-iq, and PMC-centered change -in-iq-from-time1 as my fixed factors, as well as a quadratic factor of age-from-time1 (and Iq's interactions with time). Can I simply include my IQ variables interacting with the quadratic time function, or is there something more complicated that I need to do?

Thanks!