Missing data and estimation
Posted: Mon Jan 13, 2014 2:18 am
Hello Eviews Lovers!
I'm trying to run a regression from a an excel-file i've collected at CapitalIQ, the data contains ~700 observations over M&A deals done. I've included 12 different variables and have data on most of my variables. I have both numerical like fee, speed etc. and qualitative data which i made into dummies, the dummies are only made up by one category for each variable like ( 1=yes, 0=no) so I should not end up in the Dummy Variable Trap, right?
I have some missing values for my observations, some variables are 100% covered and two that are the lowest have maybe 75% covered. The rest is pretty much covered. But when i run the regression I either get NA in my test statistics or the program says "near singular matrix - perfect coliniarity".
Does anyone know what the problem might be? When i exclude some variables it usually works. :series: :?
I'm trying to run a regression from a an excel-file i've collected at CapitalIQ, the data contains ~700 observations over M&A deals done. I've included 12 different variables and have data on most of my variables. I have both numerical like fee, speed etc. and qualitative data which i made into dummies, the dummies are only made up by one category for each variable like ( 1=yes, 0=no) so I should not end up in the Dummy Variable Trap, right?
I have some missing values for my observations, some variables are 100% covered and two that are the lowest have maybe 75% covered. The rest is pretty much covered. But when i run the regression I either get NA in my test statistics or the program says "near singular matrix - perfect coliniarity".
Does anyone know what the problem might be? When i exclude some variables it usually works. :series: :?