Feaster, et al (2011) claims contextual effects can be shown via three methods (at least): divergence of within and between effects (method 1); cross level interactions (method 2), and by estimate of the effects of discrepancies (method 3).
The variant approach to estimate the effects of discrepancies accounting for measurement error seems suited for this purposes (method 3). It would be great if varian can be expanded to suit multilevel scenarios. In which the variability within clusters can be estimated as a latent variable, to account for variance over nested outcomes.
References
Feaster, D., Brincks, A., Robbins, M., & Szapocznik, J. (2011). Multilevel Models to Identify Contextual Effects on Individual Group Member Outcomes: A Family Example. Family Process, 50(2), 167–183. doi:10.1111/j.1545-5300.2011.01353.x
Feaster, et al (2011) claims contextual effects can be shown via three methods (at least): divergence of within and between effects (method 1); cross level interactions (method 2), and by estimate of the effects of discrepancies (method 3).
The variant approach to estimate the effects of discrepancies accounting for measurement error seems suited for this purposes (method 3). It would be great if varian can be expanded to suit multilevel scenarios. In which the variability within clusters can be estimated as a latent variable, to account for variance over nested outcomes.
References
Feaster, D., Brincks, A., Robbins, M., & Szapocznik, J. (2011). Multilevel Models to Identify Contextual Effects on Individual Group Member Outcomes: A Family Example. Family Process, 50(2), 167–183. doi:10.1111/j.1545-5300.2011.01353.x