Non Normality Query
Posted: Fri Mar 13, 2015 2:01 pm
Hey guys
Hope you all are well.
I have readin many posts regarding the problem of non normality in residuals. I just have one query I was hoping you would clarify.I would very much appreciate it.
When they say non normality is not a serious problem, does this apply for classical linear regression mode as well? I'm not sure what asymtotic estimators are... but I am using the CLRM model and I want to know if I can justify in my undergrad dissertation that this assumption does not have to be corrected for.
I was hoping to use this:
(post from a couple of years ago)
A violation of the of normality assumption of the residuals is not as serious as heteroscedasticity and autocorrelation. A moderate departure from normality does not impair the conclusion when the data set is large (Bhattacharyya and Johnson 1997 p. 359). Greene (2012 pp. 64-67) states that a normal distribution of the error term is not necessary for establishing results that allow statistical inference. These results are based on the law of large numbers which concerns consistency and the central limit theorem which concerns the asymptotic distribution of the estimator.
------
But I am not sure if it's acceptable in my model. I have approximately 311 observations (monthly data for 25 years).
Also, with regards to heteroscedasticty. Is adjusting data with the White option sufficient to cancel out heteroscedasticity? Or does more need to be done. Because when I test again for heteroscedasticity, the test shows that it still exists.
Advice/guidance would be greatly appreciated. My work is due in a couple of weeks but this model building has really slowed me down!
Thanks in advance
Hope you all are well.
I have readin many posts regarding the problem of non normality in residuals. I just have one query I was hoping you would clarify.I would very much appreciate it.
When they say non normality is not a serious problem, does this apply for classical linear regression mode as well? I'm not sure what asymtotic estimators are... but I am using the CLRM model and I want to know if I can justify in my undergrad dissertation that this assumption does not have to be corrected for.
I was hoping to use this:
(post from a couple of years ago)
A violation of the of normality assumption of the residuals is not as serious as heteroscedasticity and autocorrelation. A moderate departure from normality does not impair the conclusion when the data set is large (Bhattacharyya and Johnson 1997 p. 359). Greene (2012 pp. 64-67) states that a normal distribution of the error term is not necessary for establishing results that allow statistical inference. These results are based on the law of large numbers which concerns consistency and the central limit theorem which concerns the asymptotic distribution of the estimator.
------
But I am not sure if it's acceptable in my model. I have approximately 311 observations (monthly data for 25 years).
Also, with regards to heteroscedasticty. Is adjusting data with the White option sufficient to cancel out heteroscedasticity? Or does more need to be done. Because when I test again for heteroscedasticity, the test shows that it still exists.
Advice/guidance would be greatly appreciated. My work is due in a couple of weeks but this model building has really slowed me down!
Thanks in advance