Dear all,
I had a question regarding stset for running a stcox analysis.
As a side note, I am using Stata/MP version 15.1.
I have the following long dataset (excerpt using dataex)
It tells me whether the firm failed (0/1), the predictor variables are quarterly team diversity variables (age, gender, nationality, proportion internationals).
The data should also be right censored as my data collection ended 09/12/2021, firms could have survived/failed after this date but I wouldn't know.
The variable DateEnd contains either date of failure or the last date data is collected. DateIncorporation is the date the company was founded and thus first came at risk.
I also created based on these vars the variables DaysSurvived and QuartersSurvived(days survived but scaled in quarters).
I only have info for 10 years (aka 40 quarters) of the firm as I am looking at the early stages of a firm only.
Unfortunately, after reading the stset documentation/looking online, it is still unclear if I am correctly using stset.
I was using the following stset and it produced the following outcome with no probable errors, but last observed exit t = 67 (quarters survived).
So I believe I still have to remove observations so I only look at max 40 quarters, but that's another question besides the above one.
failure event: FirmFailure != 0 & FirmFailure < .
obs. time interval: (0, QuartersSurvived]
exit on or before: failure
------------------------------------------------------------------------------
5,213,640 total observations
0 exclusions
------------------------------------------------------------------------------
5,213,640 observations remaining, representing
4,749,320 failures in single-record/single-failure data
95383520 total analysis time at risk and under observation
at risk from t = 0
earliest observed entry t = 0
last observed exit t = 67
I am wondering whether instead, I have to use this, I have to use origin and exit, though if I do this I get the following output
failure event: FirmFailure != 0 & FirmFailure < .
obs. time interval: (origin, QuartersSurvived]
exit on or before: time DateEnd
t for analysis: (time-origin)
origin: time DateIncorporation
------------------------------------------------------------------------------
5,213,640 total observations
5,213,640 observations end on or before enter()
------------------------------------------------------------------------------
0 observations remaining, representing
0 failures in single-record/single-failure data
0 total analysis time at risk and under observation
at risk from t = 0
earliest observed entry t = .
last observed exit t = .
.
Furthermore, I wondered whether it makes a difference if I use days survived vs quarters survived, and if so why?
Thank you in advance for your help.
Best regards,
Laura Hill Cabrera
I had a question regarding stset for running a stcox analysis.
As a side note, I am using Stata/MP version 15.1.
I have the following long dataset (excerpt using dataex)
It tells me whether the firm failed (0/1), the predictor variables are quarterly team diversity variables (age, gender, nationality, proportion internationals).
The data should also be right censored as my data collection ended 09/12/2021, firms could have survived/failed after this date but I wouldn't know.
The variable DateEnd contains either date of failure or the last date data is collected. DateIncorporation is the date the company was founded and thus first came at risk.
I also created based on these vars the variables DaysSurvived and QuartersSurvived(days survived but scaled in quarters).
I only have info for 10 years (aka 40 quarters) of the firm as I am looking at the early stages of a firm only.
Code:
* Example generated by -dataex-. To install: ssc install dataex clear input str16 BvdIdNumber float(DateIncorporation FirmFailure DateEnd DaysSurvived QuartersSurvived) str2 Quarters float(TMTsize BlauGender VariationAge BlauNationality PropInt) "GB11865371" 21614 1 22187 573 6 "1" 3 0 .033308666 0 0 "GB11865371" 21614 1 22187 573 6 "10" 3 0 .029356794 0 0 "GB11865371" 21614 1 22187 573 6 "11" 3 0 0 0 0 "GB11865371" 21614 1 22187 573 6 "12" 3 0 0 0 0 "GB11865371" 21614 1 22187 573 6 "13" 3 0 .028394274 0 0 "GB11865371" 21614 1 22187 573 6 "14" 3 0 .027936304 0 0 "GB11865371" 21614 1 22187 573 6 "15" 3 0 0 0 0 "GB11865371" 21614 1 22187 573 6 "16" 3 0 0 0 0 "GB11865371" 21614 1 22187 573 6 "17" 3 0 .027063293 0 0 "GB11865371" 21614 1 22187 573 6 "18" 3 0 .026646936 0 0 "GB11865371" 21614 1 22187 573 6 "19" 3 0 0 0 0 "GB11865371" 21614 1 22187 573 6 "2" 3 0 .032680206 0 0 "GB11865371" 21614 1 22187 573 6 "20" 3 0 0 0 0 "GB11865371" 21614 1 22187 573 6 "21" 3 0 .025851503 0 0 "GB11865371" 21614 1 22187 573 6 "22" 3 0 .025471335 0 0 "GB11865371" 21614 1 22187 573 6 "23" 3 0 0 0 0 "GB11865371" 21614 1 22187 573 6 "24" 3 0 0 0 0 "GB11865371" 21614 1 22187 573 6 "25" 3 0 .02474358 0 0 "GB11865371" 21614 1 22187 573 6 "26" 3 0 .02439508 0 0 "GB11865371" 21614 1 22187 573 6 "27" 3 0 0 0 0 "GB11865371" 21614 1 22187 573 6 "28" 3 0 0 0 0 "GB11865371" 21614 1 22187 573 6 "29" 3 0 .02372672 0 0 "GB11865371" 21614 1 22187 573 6 "3" 3 0 0 0 0 "GB11865371" 21614 1 22187 573 6 "30" 3 0 .02340609 0 0 "GB11865371" 21614 1 22187 573 6 "31" 3 0 0 0 0 "GB11865371" 21614 1 22187 573 6 "32" 3 0 0 0 0 "GB11865371" 21614 1 22187 573 6 "33" 3 0 .02279014 0 0 "GB11865371" 21614 1 22187 573 6 "34" 3 0 .022494167 0 0 "GB11865371" 21614 1 22187 573 6 "35" 3 0 0 0 0 "GB11865371" 21614 1 22187 573 6 "36" 3 0 0 0 0 "GB11865371" 21614 1 22187 573 6 "37" 3 0 .021924693 0 0 "GB11865371" 21614 1 22187 573 6 "38" 3 0 .021650635 0 0 "GB11865371" 21614 1 22187 573 6 "39" 3 0 0 0 0 "GB11865371" 21614 1 22187 573 6 "4" 3 0 0 0 0 "GB11865371" 21614 1 22187 573 6 "40" 3 0 0 0 0 "GB11865371" 21614 1 22187 573 6 "5" 3 0 .03149183 0 0 "GB11865371" 21614 1 22187 573 6 "6" 3 0 .03092948 0 0 "GB11865371" 21614 1 22187 573 6 "7" 3 0 0 0 0 "GB11865371" 21614 1 22187 573 6 "8" 3 0 0 0 0 "GB11865371" 21614 1 22187 573 6 "9" 3 0 .029862944 0 0 "GB11865624" 21615 0 22623 1008 11 "1" 2 .5 .1885618 .5 1 "GB11865624" 21615 0 22623 1008 11 "10" 2 .5 .19352397 .5 1 "GB11865624" 21615 0 22623 1008 11 "11" 2 .5 .19352397 .5 1 "GB11865624" 21615 0 22623 1008 11 "12" 2 .5 .19352397 .5 1 "GB11865624" 21615 0 22623 1008 11 "13" 2 .5 .1767767 .5 1 "GB11865624" 21615 0 22623 1008 11 "14" 2 .5 .1895338 .5 1 "GB11865624" 21615 0 22623 1008 11 "15" 2 .5 .1895338 .5 1 "GB11865624" 21615 0 22623 1008 11 "16" 2 .5 .1895338 .5 1 "GB11865624" 21615 0 22623 1008 11 "17" 2 .5 .173169 .5 1 "GB11865624" 21615 0 22623 1008 11 "18" 2 .5 .1857048 .5 1 "GB11865624" 21615 0 22623 1008 11 "19" 2 .5 .1857048 .5 1 "GB11865624" 21615 0 22623 1008 11 "2" 2 .5 .20203052 .5 1 "GB11865624" 21615 0 22623 1008 11 "20" 2 .5 .1857048 .5 1 "GB11865624" 21615 0 22623 1008 11 "21" 2 .5 .1697056 .5 1 "GB11865624" 21615 0 22623 1008 11 "22" 2 .5 .1820275 .5 1 "GB11865624" 21615 0 22623 1008 11 "23" 2 .5 .1820275 .5 1 "GB11865624" 21615 0 22623 1008 11 "24" 2 .5 .1820275 .5 1 "GB11865624" 21615 0 22623 1008 11 "25" 2 .5 .16637807 .5 1 "GB11865624" 21615 0 22623 1008 11 "26" 2 .5 .178493 .5 1 "GB11865624" 21615 0 22623 1008 11 "27" 2 .5 .178493 .5 1 "GB11865624" 21615 0 22623 1008 11 "28" 2 .5 .178493 .5 1 "GB11865624" 21615 0 22623 1008 11 "29" 2 .5 .16317847 .5 1 "GB11865624" 21615 0 22623 1008 11 "3" 2 .5 .20203052 .5 1 "GB11865624" 21615 0 22623 1008 11 "30" 2 .5 .1750931 .5 1 "GB11865624" 21615 0 22623 1008 11 "31" 2 .5 .1750931 .5 1 "GB11865624" 21615 0 22623 1008 11 "32" 2 .5 .1750931 .5 1 "GB11865624" 21615 0 22623 1008 11 "33" 2 .5 .16009964 .5 1 "GB11865624" 21615 0 22623 1008 11 "34" 2 .5 .17182034 .5 1 "GB11865624" 21615 0 22623 1008 11 "35" 2 .5 .17182034 .5 1 "GB11865624" 21615 0 22623 1008 11 "36" 2 .5 .17182034 .5 1 "GB11865624" 21615 0 22623 1008 11 "37" 2 .5 .15713483 .5 1 "GB11865624" 21615 0 22623 1008 11 "38" 2 .5 .16866767 .5 1 "GB11865624" 21615 0 22623 1008 11 "39" 2 .5 .16866767 .5 1 "GB11865624" 21615 0 22623 1008 11 "4" 2 .5 .20203052 .5 1 "GB11865624" 21615 0 22623 1008 11 "40" 2 .5 .16866767 .5 1 "GB11865624" 21615 0 22623 1008 11 "5" 2 .5 .18446264 .5 1 "GB11865624" 21615 0 22623 1008 11 "6" 2 .5 .1976858 .5 1 "GB11865624" 21615 0 22623 1008 11 "7" 2 .5 .1976858 .5 1 "GB11865624" 21615 0 22623 1008 11 "8" 2 .5 .1976858 .5 1 "GB11865624" 21615 0 22623 1008 11 "9" 2 .5 .1805379 .5 1 end format %td DateIncorporation
I was using the following stset and it produced the following outcome with no probable errors, but last observed exit t = 67 (quarters survived).
So I believe I still have to remove observations so I only look at max 40 quarters, but that's another question besides the above one.
Code:
stset QuartersSurvived, failure( FirmFailure )
obs. time interval: (0, QuartersSurvived]
exit on or before: failure
------------------------------------------------------------------------------
5,213,640 total observations
0 exclusions
------------------------------------------------------------------------------
5,213,640 observations remaining, representing
4,749,320 failures in single-record/single-failure data
95383520 total analysis time at risk and under observation
at risk from t = 0
earliest observed entry t = 0
last observed exit t = 67
I am wondering whether instead, I have to use this, I have to use origin and exit, though if I do this I get the following output
Code:
stset QuartersSurvived, failure( FirmFailure ) origin (DateIncorporation ) exit ( DateEnd )
obs. time interval: (origin, QuartersSurvived]
exit on or before: time DateEnd
t for analysis: (time-origin)
origin: time DateIncorporation
------------------------------------------------------------------------------
5,213,640 total observations
5,213,640 observations end on or before enter()
------------------------------------------------------------------------------
0 observations remaining, representing
0 failures in single-record/single-failure data
0 total analysis time at risk and under observation
at risk from t = 0
earliest observed entry t = .
last observed exit t = .
.
Furthermore, I wondered whether it makes a difference if I use days survived vs quarters survived, and if so why?
Thank you in advance for your help.
Best regards,
Laura Hill Cabrera
