How to generate dummy related to first value of individual observations over time

[Kodi Hannon]

Hello everybody!
I have the following dataset (reported in the table below) where an Accident can occur (when it occurs, "Accident"=1 / otherwise=0) for each ID throughout the years. What I would like to do is to generate a dummy variable (DUMMY) which takes the value of the Accident variable occurring in the first year of reference (for each ID), and then it keeps and expands that (binary) value thoughout all the remaining years (for each ID).

ID	Year	Accident	DUMMY
1	2001	0	0
1	2002	1	0
2	1999	1	1
2	2000	0	1
2	2001	1	1
3	2012	0	0
3	2013	0	0
3	2014	1	0

Here is my (unsuccessful attempt)

bysort Year: egen DUMMY = rowfirst(Accident)


Any help would be very much appreciated.
Thanks.
Kodi

[Kodi Hannon]

I have also tried:

bysort ID(Year): gen DUMMY = Accident[1]

but I don't know whether it is correct..

[Fei Wang]

Kodi, is this what you want?

* Example generated by -dataex-. For more info, type help dataex
clear
input byte id int year byte accident
1 2001 0
1 2002 1
2 1999 1
2 2000 0
2 2001 1
3 2012 0
3 2013 0
3 2014 1
end

bysort id (year): gen dummy = accident[1]

[Kodi Hannon]

Hello Fei! Nice cat
hum.. I think the sum function doesn't work.

[Fei Wang]

Hello Fei! Nice cat
hum.. I think the sum function doesn't work.

I misunderstood your request, and your solution in #2 is correct.

[Carlo Lazzaro]

Kodi:
do you mean something along the following lines?

. bysort ID ( Year ): gen wanted=1 if DUMMY==1 & _n==1

. bysort ID ( Year ): replace wanted=wanted[1] if wanted[1]==1

. replace wanted=0 if wanted==. & DUMMY!=.

. list

     +---------------------------------------+
     | ID   Year   Accident   DUMMY   wanted |
     |---------------------------------------|
  1. |  1   2001          0       0        0 |
  2. |  1   2002          1       0        0 |
  3. |  2   1999          1       1        1 |
  4. |  2   2000          0       1        1 |
  5. |  2   2001          1       1        1 |
     |---------------------------------------|
  6. |  3   2012          0       0        0 |
  7. |  3   2013          0       0        0 |
  8. |  3   2014          1       0        0 |
     +---------------------------------------+

.

[Kodi Hannon]

Buongiorno Carlo.
Grazi. Si!
Kodi