I would like to add control variables (BookToMarket, Size...) to my data to regress Abnormal return on text sentiment. For each firm and specific date, I have several thousands of row of data.
Each row is a unique sentence and variables associated with it.
I have included a sample of my data.
If I add the size(log of market cap), BooktoMarket ...., then I would be adding the same values for thousands of rows which raises some question marks on my mind. If my data was based on days rather than intraday data, I would have included the values for those variables.
I can also tag each of those values according to quintiles (1...5) and then use it in my regression as categorical variable.
In the sample data, I categorized variables (ln_mktcap bm roa) according to quintiles.
Should I just use the raw data rather than quintiles in my regression? What would be the best way to add those variables to my regression?
reghdfe AbnRet i.Sentiment i.ln_mktcap i.bm i.roa, absorb(year SIC) cluster(month_year)
Thank you.
Code:
* Example generated by -dataex-. For more info, type help dataex clear input byte Sentiment double AbnRet float(year month_year) byte(SIC ln_mktcap bm roa) 0 .000042760626015603265 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 0 -.00008552490912983046 2017 685 38 5 3 5 0 .0000427551413056948 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 0 -.0010592237732415244 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 1 -.00004275514130580582 2017 685 38 5 3 5 1 0 2017 685 38 5 3 5 1 -.00004275331338177146 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 1 0 2017 685 38 5 3 5 1 -.000042751485614056506 2017 685 38 5 3 5 1 0 2017 685 38 5 3 5 1 0 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 1 0 2017 685 38 5 3 5 0 .00004274965800266095 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 1 0 2017 685 38 5 3 5 1 0 2017 685 38 5 3 5 1 0 2017 685 38 5 3 5 1 0 2017 685 38 5 3 5 2 0 2017 685 38 5 3 5 1 .000042751485614056506 2017 685 38 5 3 5 1 .000042753313381882485 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 2 0 2017 685 38 5 3 5 1 0 2017 685 38 5 3 5 2 0 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 1 0 2017 685 38 5 3 5 2 0 2017 685 38 5 3 5 1 0 2017 685 38 5 3 5 1 .0000427551413056948 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 1 0 2017 685 38 5 3 5 1 0 2017 685 38 5 3 5 1 0 2017 685 38 5 3 5 1 0 2017 685 38 5 3 5 1 0 2017 685 38 5 3 5 1 0 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 0 -.00012828187804658775 2017 685 38 5 3 5 0 .0001282654239170844 2017 685 38 5 3 5 0 .0021446675765255385 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 0 .000042760626015603265 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 1 0 2017 685 38 5 3 5 1 0 2017 685 38 5 3 5 1 0 2017 685 38 5 3 5 1 0 2017 685 38 5 3 5 1 0 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 1 0 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 1 0 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 1 -.0001282873636947457 2017 685 38 5 3 5 0 .0000855139387719861 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 1 0 2017 685 38 5 3 5 1 0 2017 685 38 5 3 5 2 0 2017 685 38 5 3 5 2 0 2017 685 38 5 3 5 1 -.00008552125203098448 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 1 0 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 1 .0000855139387719861 2017 685 38 5 3 5 1 1.1102230246251565e-16 2017 685 38 5 3 5 1 0 2017 685 38 5 3 5 1 0 2017 685 38 5 3 5 0 .00008552125203120653 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 0 0 2017 685 38 5 3 5 2 0 2017 685 38 5 3 5 1 -.0000427642832705466 2017 685 38 5 3 5 1 0 2017 685 38 5 3 5 1 0 2017 685 38 5 3 5 0 .00004276245456480421 2017 685 38 5 3 5 end format %tm month_year label values Sentiment sentimentlabel label def sentimentlabel 0 "Neu", modify label def sentimentlabel 1 "Pos", modify label def sentimentlabel 2 "Neg", modify
0 Response to Control variables for intraday data.
Post a Comment