Hello everyone! My research model includes both Level 1 (individual) and Level 2 (team) variables. Initially, the goal is to explore the determinants of innovative work behavior at the individual level, which is why I ran a multilevel analysis in Stata. The second part of the research has as dependent variables two types of innovation at the team level (exploratory and exploitative innovation).
Wanting to run an overall model I employed Gsem in stata 14. I understand that the second part of the research does not require multilevel analysis since my dependent variables are at the upper level (level 2). In order to connect the two strands of research I aggregate innovative work behaviour to the team level (after checking for rwg, ICC1,ICC2). But is this correct ? In other words, beyond aggregation how can I regress exploratory and exploitative innovation from a level 1 variable-here employee's innovative work behaviour-? Does anyone employed the latent variable approach proposed by Croon et al (2007)* with GSEM ? Below I attach the gsem model that I specify.
PS My sample consists of 299 employees nested within 103 teams (~3 people per team).
Your help is much appreciated.
*Croon, M. A., & van Veldhoven, M. J. P. M. (2007). Predicting grouplevel outcome variables from variables measured at the individual level: A latent variable multilevel model. Psychological Methods,Array 12, 45–57.
0 Response to Model Specification with GSEM Stata including both multilevel and simple regression modelling?
Post a Comment