Page 1 of 1
non normal distribution
Posted: Mon May 06, 2013 4:12 am
by rahma
Hello :)
The temporal serie is not normal distributed, even when i enter the Log
What solution do you suggest to me?
Re: non normal distribution
Posted: Mon May 06, 2013 6:59 am
by startz
What exactly is the problem with the series being non-normal ?
Re: non normal distribution
Posted: Mon May 06, 2013 8:18 am
by rahma
for the cointegration test, i have 3 temporal series and they are all non normal, it does not matter»?
Re: non normal distribution
Posted: Mon May 06, 2013 8:35 am
by startz
The series being normal is not of any relevance for cointegration tests. Normality of series is almost never important. Normality of errors sometimes matters, but not often.
Re: non normal distribution
Posted: Mon May 06, 2013 8:38 am
by rahma
ok, thank you :)
Re: non normal distribution
Posted: Mon May 27, 2013 9:45 am
by Cabinho
Hi,
Is non normal distribution in cross sectional data a problem?
If it is still existent even after logging data, what can be done?
Thank you
Re: non normal distribution
Posted: Mon May 27, 2013 10:05 am
by startz
Non-normality of data is rarely a problem in large samples unless the departure from normality is extreme.