Assuming a regression equation y=a+bx+u. Where, the point of interest is the coefficient of x "b", which is required to regress on some other variable, let's say "z". In this case, can we generate a series for the beta as "b= (y-a-u)/x ", is it statistically/econometrically appropriate to carry on with this? or is there any other way to approach this issue.
Looking forward to your advice.
Well, if you just consider the point estimate it is a constant so you just have b=b for your series. if you consider b=(y-a-u)/x then everything is a variable and two are random. How would you expand that? I don't have any idea what you are trying to do. you might look at errors in variables models. I would start at the link below. then follow with the files. Best, David
https://en.wikipedia.org/wiki/Errors-in-variables_models
We can not generate a series for the beta as "b= (y-a-u)/x " for a linear regression equation y=a+bx+u.
Professor David Eugene Booth thank you for your time and concern. I quite this idea as it does not serve my purpose.
- Badges
- Science topic
- Similar topics
- Condensed Matter Physics
- Zero