[R-sig-ME] Small data set and validity of Multilevel Regression results

[Linus Holtermann]

Hello,

I want to estimate a hierarchical multilevel regression (two levels). My data set is fairly small: 25 groups and 255 individual observations distributed unequally among groups (group member size varies between 4 and 55). Several books point out that the small sample size could lead to biased estimates of the random effects. There was no clear conclusion which reasons are responsible for the bias. 
My question is as follows: which of the statements are true?
-  Adding (more) regressors to Level 1 (individuals) will reduce the validity of the estimation results
- Adding (more) cross-level intaraction effects will reduce the validity of the estimation results
- Adding (more) random slopes will reduce the validity of the estimation results

Could a MCMC approach help in my example? 

Thanks in advance, 



 
Linus Holtermann
Hamburgisches WeltWirtschaftsInstitut gemeinnützige GmbH (HWWI)
Heimhuder Straße 71
20148 Hamburg
Tel +49-(0)40-340576-336
Fax+49-(0)40-340576-776
Internet: www.hwwi.org
Email: holtermann at hwwi.org
 
AmtsgerichtHamburg HRB 94303
Geschäftsführer: Prof. Dr. Thomas Straubhaar, Gunnar Geyer
Umsatzsteuer-ID: DE 241849425