Result viewer space falls short

[ajay pasi]

Hi statalist community, I am running a long regression with almost 400000 lines (due to fixed effect in cross sectional data). But after running the regression, and scrolling up, the space falls short, and I am not able to scroll further up, to see the regression result.

regards,
ajay

[Clyde Schechter]

I am running a long regression with almost 400000 lines (due to fixed effect in cross sectional data).

What is the exact code you are using for this regression? If you are emulating a fixed-effects regression using -regress ... i.fixed_effect...-, you can get output that suppresses the output of fixed-effect coefficients by using -xtreg, fe-, -areg-, or Sergio Correa's -reghdfe- (available from SSC) instead. See the -help- files for those commands if you are not familiar with them.

[ajay pasi]

Prof. Clyde, thanks for the help. I was running 'logit' with i.fixed_effect.

ajay

[ajay pasi]

Prof. Clyde, one more thing. I installed it, can I use Sergio Correa's module for logit analysis?

[ajay pasi]

I guess not! could you please suggest a module like Sergio Correa's that I can use for logit.

[Clyde Schechter]

-xtlogit, fe- does that. I'm not aware of anything analogous to -reghdfe- for logistic regression.

I should also add that although the Results window has a scroll buffer whereby eventually things "fall off" the top when output gets long, if you are logging your work as you go (which you should be doing anyway so you have an audit trail of what you have done), the resulting log file can be read in the viewer, and its not limited the way the Results window is. So

capture log close
log using name_of_desired_logfile, replace

// PRODUCTION CODE GOES HERE

log close

To see the results:

view name_of_desired_logfile.smcl

That said, scrolling through 400,000 lines in the log file, much of which is irrelevant, is probably not fun either.

[ajay pasi]

Thank you sir. I get it.

[ajay pasi]

Sir, sorry for bothering you one more time.

If I need many fixed effects in xtlogit e.g., xtlogit depVar indepVar i.fixed_effect_1 i.fixed_effect_2 i.fixed_effect_3, fe cluster (clusterVar_1)

Then before running above mentioned xtlogit............sterVar_2), should I run-- xtset fixed_effect_1 fixed_effect_2 fixed_effect_3. Am I right in doing that?

[Clyde Schechter]

Then before running above mentioned xtlogit............sterVar_2), should I run-- xtset fixed_effect_1 fixed_effect_2 fixed_effect_3. Am I right in doing that?

No, -xtset- only allows you to specify one panel variable, and, optionally, one within-panel variable. (Typically, these are the unit of analysis such as person or firm, and time.) And -xtlogit- will only absorb the fixed effects for the panel variable (N.B. not the within-panel variable). If you have additional fixed effects, they have to be spelled out as i.varname.

[ajay pasi]

I get it. Appreciate your help. Sorry for my childish questions!