Growth Curve Episode 7: Time-Varying Covariates

[@johngalt5258]

These videos are incredibly helpful -- thanks so much for making them available.

[@cutepanda666]

Patrick, I am truly grateful for your passion and talent for making something seemingly impossible into comprehensible and manageable concepts. Big hug

[@centerstat]

Jaime -- Thanks for your very kind words. I'm sincerely glad you find these of some use. Stay safe -- patrick

[@Davao420]

I love you guys! I so need this for my PhD

[@donaldschoolmasterjr9423]

Amazingly clear explanation!

[@장동일-b6s]

Thanks! This lecture helps me a lot.

[@goelnikhils]

Amazing video on TVC

[@Asdf-we7vr]

wow... Thank you so much for this short lecture

[@realjcluo]

Looking forward to the introduction of the multivariate latent curve model!!!

[@emmaburns1758]

Thank you so much for these videos! They are so helpful. I had a follow up question for LGM with TVCs. Is it possible to examine a factor (for example self-efficacy) as both a time variant and time invariant factor within a single model? Such that the first time point would be used to assess the effects on the slope and intercept but also its contemporaneous effects?

[@mcClinas1]

Hi Professor Curran. Thanks again for creating these videos. You make a very complicated topic both intelligible and interesting. I have a question about time-variant covariates (TVCs). Because variables tend to be highly correlated over time (e.g., alcohol use at Time 1 may be highly correlated with alcohol use at Time 2), is there any concern of multicolinearity with TVCs? I have found, in my own data, that a TVC predicts an outcome, but it appears that effect is due a high correlation among the repeated measures of the TVC. Thanks again and hope you create more videos! I'm a fan. -Andrew

[@mcClinas1]

Thank you Dr. Curran!!! -Andrew

[@amypacker3032]

Thanks so much for sharing these helpful videos! TVCs questions: if you are including time-varying covariates in an unconditional model, should these be included when determining the shape of the curve? Or should you determine the shape first, using fit indices and LRT tests, before adding unconditional time-varying covariates?

[@RuhulAmin-th3zj]

thanks a lot sir

[@annan.9000]

Dear Dr. Curran, Thanks so much for these videos and for the awesome Quantitude podcasts. I'm a fan! At 10:47 into the video, you mentioned that it's standard to covary time-varying covariates with the slope factors. My questions are: when do you NOT want to covary the time-varying covariates with the slope factors and how does that change the interpretatations regarding the effect of time-varying predictors on the outcome variables? Right now, when they're correlated, the correct interpretation in this video is the effect of environmental stress on the antisocial behavior beyond the growth factor (i.e., slope). How does the interpretation change when TVCs are not correlated with slope? I ask this question because I try to run a similar model in Mplus and by default, it doesn't automatically correlate my 3 slopes (3 because I'm running a piecewise growth model) with TVCs. Thank you for any guidance you can provide.

[@centerstat]

Hi Anna -- thanks for your nice note. Briefly, if the TVCs and growth factors are correlated then the TVC is a pure estimate of the within-person component of the TVC to outcome relation. However, if these are uncorrelated, then the TVC is an aggregate of the within-person and between-person components of the TVC to outcome relation. We puzzle through these issues in the paper below. Good luck with your work -- patrick Curran, P.J., Lee, T.H., Howard, A.L., Lane, S.T., & MacCallum, R.C. (2012). Disaggregating within-person and between-person effects in multilevel and structural equation growth models. In J. Harring & G. Hancock (Eds.) Advances in Longitudinal Methods in the Social and Behavioral Sciences (pp. 217-253). Charlotte, NC: Information Age Publishing.

[@annan.9000]

@@centerstat Ah I see. Thank you so much, Dr. Curran! I will check out the chapter you provided.

[@fazlfazl2346]

Wonderful lecture again: I am really confused by the interpretation of the interaction between time and another time varying covariate e.g. Cholesterol*time in growth curve and longitudinal models. How do we interpret Cholesterol*time 1) when only time has a random slope 2) when both Cholesterol and time have random terms for slope?

[@centerstat]

Thanks for the kind words -- the interaction between time and a TVC would test whether the magnitude of the effect of the TVC in the prediction of the outcome varies systematically with the passage of time. In other words, might the magnitude of the TVC systematically strengthen or weaken as time progresses.

[@user-tt2hp5zx3m]

Thank you for these videos! Just a quick question. If you have a time-varying covariate that changes systematically over time (e.g., fits a random effects linear/quadratic model) do you HAVE TO run a multivariate multilevel model or SEM so all sources of variances are accounted for, or can you still run a time-lagged or standard MLM and include these as covariates with these variables? Thank you!! Holly

[@centerstat]

Hi Holly -- thanks for your note. This one is tricky -- it's not so much that the TVCs are growing and and of themselves, but it's how you person-mean center to isolate within vs between person effects. That is, if your TVC is systematically growing over time (whatever form that might be -- linear, quadratic, etc.) then do the simple person-mean centering will give you a biased estimate of your within-person effects (because the mean assumes no growth over time). I've written a couple of things on this, as have several others -- a few cites are below. Good luck with your work -- patrick Curran, P.J., & Bauer, D.J. (2011). The disaggregation of within-person and between-person effects in longitudinal models of change. Annual Review of Psychology, 62, 583-619. Curran, P.J., Lee, T.H., Howard, A.L., Lane, S.T., & MacCallum, R.C. (2012). Disaggregating within-person and between-person effects in multilevel and structural equation growth models. In J. Harring & G. Hancock (Eds.) Advances in Longitudinal Methods in the Social and Behavioral Sciences (pp. 217-253). Charlotte, NC: Information Age Publishing. Curran, P.J., Howard, A.L., Bainter, S.A., Lane, S.T., & McGinley J.S. (2013). The separation of between-person and within-person components of individual change over time: A latent curve model with structured residuals. Journal of Consulting and Clinical Psychology, 82, 879-894. Hamaker, E. L., Kuiper, R. M., & Grasman, R. P. (2015). A critique of the cross-lagged panel model. Psychological Methods, 20, 102. Wang, L. P., & Maxwell, S. E. (2015). On disaggregating between-person and within-person effects with longitudinal data using multilevel models. Psychological Methods, 20, 63.