I'm trying to find the effect of smart cities on sustainable development. I collected data for several indicators of smartness of a city for 15 cities over 12 years. seems like panel data (GMM methods) is not appropriate considering the small sample size. Another method could be structural equation modelling which I'm not sure if it's appropriate for panel data. any suggestions for SEM or any other appropriate method?
Use this book:
Longitudinal Data Analysis Using Structural Equation Models
Jack J. McArdle and John R. Nesselroade
Tara Vanlı, yes, you can use the structural equation modelling (SEM) framework for analysing panel data. Here is a paper illustrating how this can/should be done: Article Linear Dynamic Panel-data Estimation Using Maximum Likelihoo...
Dear Tara Vanlı,
Kuklys (2005) use a panel survey to calculate index in each dimension and then create a second index that "combines" the dimensions. He uses SEM to estimate each index (of education and housing) and to combine them. It is worth looking at how he modeled those index. You can do something like that in STATA (use sem or gsem in STATA), simplify the structural equation and give more attention to the measurement equation. See Kuklys (2005) "Amartya Sen's capability Approach".
- Badges
- Science topic
- Similar topics
- Thesis Research