*Now* do you see what I mean? C On Wed, 26 Jan 2022 at 14:30, Sven Schreiber <svetosch(a)gmx.net&gt; wrote: -- Clive Nicholas "My colleagues in the social sciences talk a great deal about methodology. I prefer to call it style." -- Freeman J. Dyson

In my text recommendations, I intended to say that Wooldridge is an introductory econometrics book that contains a good coverage of panel data econometrics. Note the following 1. Unobserved heterogeneity is fundamental to panel data econometrics. Standard regression can not account for unobserved heterogeneity and therefore you will have a bias in your estimates. 2. The choice of fixed or random effects depends on the existence of correlation between the unobserved effects and the explanatory variables. 3. There are panel data estimates that are robust with respect to autocorrelation (and heteroskedasticity). There are also panel methods for dynamic models which will not be covered in introductory texts. 4. Multicollinearity is mainly a data problem. As far as I know, there are no multicollinearity routines specific to panel data. This is a separate topic. 5. I would think of non-normality as indicative of some model specification error. A lot of this is covered in chapters 13 and 14 of the Wooldridge introductory text. If you find this text too mathematical can I suggest that you seek out an econometrician who is willing to collaborate with you? Otherwise, you need to put considerable effort into upgrading both your mathematics and your econometrics. There are fora such as StackExchange, Quora, or ResearchGate where you can get help on problems relating to Econometric/statistical/mathematical theory as opposed to the implementation of such items in gretl. John C Frain 3 Aranleigh Park Rathfarnham Dublin 14 Ireland www.tcd.ie/Economics/staff/frainj/home.html https://jcfrain.wordpress.com/ https://jcfraincv19.wordpress.com/ mailto:frainj@tcd.ie mailto:frainj@gmail.com On Wed, 26 Jan 2022 at 15:33, Clive Nicholas <clivelists(a)googlemail.com&gt; wrote: