There are many reasons ...check for larger values or outliers or is it a qualitative data or use the test which don't ask for normality as an assumption or do two way transformation. If you still have problems do respond
@@dr.shobhak6764 mam when even after doing log or squaring the values, data is not coming normal, then what are the options? Can We still go for parametric tests? Pl share your email.
@@dr.shobhak6764 Thanks Dr! I am facing the same problem, please what does the two way transformation mean? And what is the test that does not ask for normality? Can we analyze with a series that does not follow a normal law and why Dr? Glad to read your answer.
@@dr.shobhak6764 Ma'am thank you for you kind reply. I have another question, if one of my variable is not stationary in a model then what should I do? Kindly help me in this regards
@@rafeeqahmed941 if certain variables are stationary at level and certain at first difference use ARDL. If the variables are stationary at second difference then you have to opt for Toda Yamamoto test
Initially u have converted my all variable in log form and run further analysis and when at last stage when I perform VECM diagnostic test residuals seem to be non normal How to solve this problems
Ma'am my balance of payment data is in negative. I have collected it from RBI. how to convert it in log form. When I converted it, it is showing as missing. But the data is already in negative form given by RBI. what to do here. Kindly give your suggestion
what if the log of the data does not solve the non normality test, is there any other way of making the data normally distributed?
i have the same issue
Thanks, Professor.
What if i have a large sample size where data is not normally distributed in both forms original and log form data?
thank you so much!
Thank you
if after i converted to log and the p value (less 5%) of Jacque Berra test does not accept Ho. how solve the problem?
There are many reasons ...check for larger values or outliers or is it a qualitative data or use the test which don't ask for normality as an assumption or do two way transformation. If you still have problems do respond
@@dr.shobhak6764 mam when even after doing log or squaring the values, data is not coming normal, then what are the options? Can We still go for parametric tests? Pl share your email.
shobhagacecocbe@gmail.com
Check out for quantile regression
@@dr.shobhak6764 Thanks Dr! I am facing the same problem, please what does the two way transformation mean? And what is the test that does not ask for normality? Can we analyze with a series that does not follow a normal law and why Dr? Glad to read your answer.
Ma'am if there are 7 variables and yearly data of 53 years can I use ARDL MODEL? ANOTHER THING I HVAE ONE DEPENDENT VARIABLE AND OTHER ARE REGRESSORS.
Just check for the conditions before using ARDL
@@dr.shobhak6764
Ma'am thank you for you kind reply.
I have another question, if one of my variable is not stationary in a model then what should I do? Kindly help me in this regards
@@rafeeqahmed941 if certain variables are stationary at level and certain at first difference use ARDL. If the variables are stationary at second difference then you have to opt for Toda Yamamoto test
Initially u have converted my all variable in log form and run further analysis and when at last stage when I perform VECM diagnostic test residuals seem to be non normal
How to solve this problems
Check for outliers.
@@dr.shobhak6764 I also checked it ma'am
There is no outliers in this series
You should go for structural break. I haven't worked on it.
@@dr.shobhak6764 Dr how to know the aberanters values? Thanks for your answerr.
Thank you mam😊
Love u mam.
Ma'am my balance of payment data is in negative. I have collected it from RBI. how to convert it in log form. When I converted it, it is showing as missing. But the data is already in negative form given by RBI. what to do here. Kindly give your suggestion
Leave that particular year. If it's negative you can't convert to log.
@@dr.shobhak6764 Thank You So much Ma'am. Sending Gratitude.
my question is also if the data after log transformation is not normal then what should to do?
Quantile regression
Check my channel
what is prob is completely zero
If p is 0 then its significant at 1 percent level