Susana Bird
2003-Sep-22 11:45 UTC
[R] weighting the OLS regression by the inverse of the number of observations
Dear R-proffecionals and users, I have the following two questions. I have the cross-country sample. The obeservaions in each country varie substantially and could not use it for the regression because I have several outliers. I do not want to lose any information from the sample and would like to use it. I would like to include the all of countries into regression analysis. Yet, I have the problem. I learned that I should in my case use the weight as an inverse to the observation’s number in each country. I am not sure if it is right and how I should compute it by using R. In addition: coudl I run the ordered probit regression by this weight? It was the first question. The second question is the following. I found the DFBETA-Test, that should be very usefull for the purpose of my study, but I could not find any literature how to use it. Do anybody know how to use it by using R and where could I find any information more? Thanks a lot in advance, Susana --------------------------------- Yahoo! Search - Looking for more? Try the new Yahoo! Search [[alternative HTML version deleted]]