I collected longitudinal data with irregular time intervals. I collected data from mobile app users in three waves of surveys soon after they finished using the app. Can you suggest the best method for analysing longitudinal data with uneven time interval? and the software which can be used for such analysis?
A common approach to longitudinal designs is treating measurement as nested within the person and using time (however measured) as a covariate. Google multilevel and longitudinal.
Hi Mona,
you could take a look onto growth curve modeling, where uneven time intervals can be represented by unevenly progressing "factor loadings" of a latent growth factor.
Actually, Daniel Wright 's suggestion of a multilevel approach is just a different representation for this analysis and will result in almost identical results. The only difference is that a multilevel model assumes equal error variances of the measures across time and growth curve modeling does not (and you can test that).
Here's some refs
Biesanz, J. C., Deeb-Sossa, N., Papadakis, A. A., Bollen, K. A., & Curran, P. J. (2004). The role of coding time in estimating and interpreting growth curve models. Psychological Methods, 9(1), 30-52.
Bliese, P. D., Ployhart, R. E., & Poyhart, R. E. (2002). Growth modeling using random coefficient models: Model building, testing, and illustrations. Organizational Research Methods, 5(4), 362-387. doi:10.1177/109442802237116
Bollen, K. A., & Curran, P. J. (2004). Autoregressive latent trajectory (ALT) models A synthesis of two traditions. Sociological Methods & Research, 32(3), 336-383. doi:10.1177/0049124103260222
Garst, H., Frese, M., & Molenaar, P. C. M. (2000). The temporal factor of change in stressor-strain relationships: A growth curve model on a longitudinal study in east germany. Journal of Applied Psychology, 85(3), 417-438.
Hox, J., & Stoel, R. D. (2005). Multilevel and SE>M approaches to growth curve modeling. In B. S. Everitt & D. C. Howell (Eds.), (Vol. 3, pp. 1296-1305). Chichester: John Wiley & Sons.
Li, F., Duncan, T. E., & Acock, A. (2000). Modeling interaction effects in latent growth curve models. Structural Equation Modeling, 7(4), 497-533.
Mehta, P. D., & West, S. G. (2000). Putting the individual back into individual growth curves. Psychological Methods, 5(1), 23-43.
And here is an open available workshop on SEM with mplus with videos. The last 4 sessions are on longitudinal analyses:
https://bit.ly/2VCW7pK
Best,
Holger
Just following up from Holger Steinmetz 's detailed and excellent answer, if you conceive of the random variables as latent, then conceptually mixed/multilevel models can be used for similarly to problems as latent variable ones, though sometimes the computations differ so it takes some care to get the same results (I used to show this with the latent growth curve lecture). And heterscedaticity can be included in either.
You need to do some readings in Longitudinal design and also multilevel analysis
- Badges
- Science topic