What can you possibly be thinking of? Best wishes David Booth

Thank you Prof. David Eugene Booth for your prompt response.

I want to use a quantile regression based on three-group index (low, medium, and high) to assess how variable (X) impacted variable (Y) in each period using ARIMA. In Stata, the commands used for ARIMA and quantile regression separately are known, but the issue is how to perform ARIMA for each period of the three-group index as mentioned above.

Let's try this another way. What is your research question? Thanks, David Booth

You can try the below commands:

"arima consump m2 if tin (, low), ar(1) ma(1)".

"arima consump m2 if tin (, medium), ar(1) ma(1)".

"arima consump m2 if tin (, high), ar(1) ma(1)".

Instead of consumption and m2 you use those variables suitable for your research.

Please read the Stata command file: https://www.stata.com/features/time-series/ts-arima.pdf

Especially Example 4.

Dear Prof. David Eugene Booth,

How is the impact of the X on Y during

uncertainty periods (low, medium, high).

Thanks

Dear Dr. Faruque As Sunny

Thanks for your help.

The command "tin" is suitable for sub-samples, for example differnt years, etc. In my case, how to add the variable (uncertainty periods) that should be divided into groups (low, medium, high) to the equation using "tin" command?. I thinks there is another command that should be used.

Thanks

Please try to code low=1, medium=2, high=3 or any number you prefer and try. Hope this will solve the problem.

I agree with you. I have done something like that as a solution but still looking for the command that can do it directly.

Thanks again Dr. Faruque As Sunny

Welcome Dr. Naji Mansour Nomran. If you still have confusion, please post in Statelist.org. I think you will get better suggestions.

Good suggestion Dr. Faruque As Sunny.

Best regards