How to add a new column in stata by using a vector and how to create a column by using results from regression?

[Jiali Luo]

1. I am coding in a do file, and want to add a column of year in stata , ranging from year 1989 to year 2015. May I ask the code to do it?
2. I also want to create a column for adjusted r square from my regression model (regressed yearly from year 1988 to 2015). That is to say, 27 observations of adjusted r square. May I ask the code for this purpose?

[Nick Cox]

It would be best to tell us first about your existing data.

[Jiali Luo]

It would be best to tell us first about your existing data.

Hi Nick. I am thrilled to your reply. I didn't come back to Statlist for a while cuase I thought my silly question wouldn't get any attention.

Frist, I attach the table I wanna replicate for you to better understand what I am trying to do.

Second, to answer your question. I have the data of firms' cash flow and earnings from year 1989 to year 2015. I want to extract the adjusted R square for xtreg CF_TA l.EARN_TA for each year and then list them in the database in the order of year as illusted in my attachement.

For example, this is how I display the adjusted of R square for one year. There are two issues with my method: first, I don't know how to put the R square value in the database in the year of order; Second, I don't know how to do a loop here. If I want to extract all the adjusted R square for 27 years, I will have to repeat the following code 27 times.

xtreg CF_TA l.EARN_TA if (fyear == 1990)
ereturn list
scalar rsquare = e(mss)/(e(mss)+e(rss))
scalar r2_a = 1 - (1 - e(r2))*( e(df_r) + e(df_m) )/( e(df_r) )
display r2_a

Please let me know if my expression is still unclear.

[Ken Chui]

One possible way is to turn the analysis code into a program, and then use -statsby- to apply that across years. Since you didn't use -dataex- to provide any sample data, here is one that uses the built-in auto data set as an illustration:

sysuse auto, clear

capture program drop regProg
program define regProg, rclass
    reg mpg weight headroom
    * Whatever you'll need to compute:
    return scalar smthng = e(r2_a)
end

* Check it out:
regProg
return list

* Apply across different levels in rep78:
statsby adj_r2 = r(smthng), by(rep78): regProg

The existing data will be replaced by the requested summary statistics.

[Jiali Luo]

Hi Nick. I am thrilled to your reply. I didn't come back to Statlist for a while cuase I thought my silly question wouldn't get any attention.

Frist, I attach the table I wanna replicate for you to better understand what I am trying to do.

Second, to answer your question. I have the data of firms' cash flow and earnings from year 1989 to year 2015. I want to extract the adjusted R square for xtreg CF_TA l.EARN_TA for each year and then list them in the database in the order of year as illusted in my attachement.

For example, this is how I display the adjusted of R square for one year. There are two issues with my method: first, I don't know how to put the R square value in the database in the year of order; Second, I don't know how to do a loop here. If I want to extract all the adjusted R square for 27 years, I will have to repeat the following code 27 times.

xtreg CF_TA l.EARN_TA if (fyear == 1990)
ereturn list
scalar rsquare = e(mss)/(e(mss)+e(rss))
scalar r2_a = 1 - (1 - e(r2))*( e(df_r) + e(df_m) )/( e(df_r) )
display r2_a

Please let me know if my expression is still unclear.

Hi Ken, Thank you for the code. It is very helpful. I am sorry that I didn't know I could use -dataex- and didn't provide the data. I will definitely try -dataext- next time