Hi Miguel, you might want to have a look at an old working paper version of my paper 'Explaining fixed effects: random effects modelling of TSCS and Panel data'.

http://polmeth.wustl.edu/media/Paper/FixedversusRandom_1_2.pdf

That has an example on income using panel data, and includes Stata code for the models we fitted. Unfortunately we had to take that example out of the published version because it was too long. The published version does include the other example (in an appendix) and can be found here: http://journals.cambridge.org/action/displayAbstract?fromPage=online&aid=9251648&fileId=S2049847014000077 or on my researchgate profile.

Hope that helps!

Thank you very much Andrew, I think that this article will be very useful, and i will try to see also the example.  When I get to read it in detail I will write to you.  Thank you very much for your help...  ! 

You might also want to have a look at this which is freely available on RGate - it has some  long chapters on the random effects model for analysing repeated measures data

https://www.researchgate.net/publication/260772180_Developing_multilevel_models_for_analysing_contextuality_heterogeneity_and_change_using_MLwiN_Volume_2

Book Developing multilevel models for analysing contextuality, he...

Thank you very much Kelvin! I find this article didactic and clear for non statisticians, like me (sociologist with some experience in quantitative research and applied statistics).  I have a little question, in figure 6 (p. 106) case c) looks like case b) in table 7, or have i got it wrong? It says descreasing variability, and in figure 6 it looks like increasing.  Best regards and I thank you again for this excelent and clear text... Miguel

Thanks for the supportive comments - you are correct the graphs and text need switching - b) is decreasing c) is increasing heterogeneity. I have uploaded a corrected version.