Identifying variable that contains rowmax value

[Al Perez]

I am trying to identify the variable that equals the rowmax variable, but am running into issues.

I have school district-level data that contains three variables indicating the percentage of students in the district receiving a certain instruction type: remote, in-person and hybrid. Each row is a unique district. I do the following:

egen max = rowmax(in_person remote hybrid)
gen inst_type = ""
replace inst_typ = "in_person" if max == in_person
replace inst_typ = "hybrid" if max == hybrid
replace inst_typ = "remote" if max == remote

Here's a table to illustrate:

district_id	remote_pct	hybrid_pct	in_person_pct	max	inst_type
1	88.5	11.5	0	88.5	remote
2	11.5	77.8	5.1	77.8	""
3	94.4	0	5.6	94.4	""
4	23.1	74.1	2.8	74.1	""

The variable max is correctly generated. The problem is that inst_type isn't modified the way I want. For each district, I am trying to indicate in which instruction mode most students participated. For example a district for which max==remote will indeed have the correct values but inst_type will be missing instead of inst_typ == "remote". What could I be doing wrong? Apologies if this is a rudimentary question, but I've been at this for over an hour. I am using STATA 16.

Thanks for the help.

[Ken Chui]

I used the following codes:

clear
input district_id     remote_pct     hybrid_pct     in_person_pct     max     str30 inst_type
1     88.5     11.5     0     88.5     remote
2     11.5     77.8     5.1     77.8     ""
3     94.4     0     5.6     94.4     ""
4     23.1     74.1     2.8     74.1     ""
end
list

drop max inst_*

egen max = rowmax(in_person remote hybrid)
gen inst_type = ""
replace inst_typ = "in_person" if max == in_person
replace inst_typ = "hybrid" if max == hybrid
replace inst_typ = "remote" if max == remote

list, sep(0)

And I was not able to replicate your errors. This is what I got:

     +-------------------------------------------------------------+
     | distri~d   remote~t   hybrid~t   in_per~t    max   inst_t~e |
     |-------------------------------------------------------------|
  1. |        1       88.5       11.5          0   88.5     remote |
  2. |        2       11.5       77.8        5.1   77.8     hybrid |
  3. |        3       94.4          0        5.6   94.4     remote |
  4. |        4       23.1       74.1        2.8   74.1     hybrid |
     +-------------------------------------------------------------+

Consider using -dataex- to post some sample data that can let us reproduce that error.

[Ali Atia]

Try:

egen double max = rowmax(in_person remote hybrid)

which should get you the results you're looking for.

See also -help precision- for more information why you are encountering these problems.

[Al Perez]

Try:

egen double max = rowmax(in_person remote hybrid)

which should get you the results you're looking for.

See also -help precision- for more information why you are encountering these problems.

Thank you! This was exactly the issue. I had a feeling it had something to do with the format...

[Al Perez]

I used the following codes:

clear
input district_id remote_pct hybrid_pct in_person_pct max str30 inst_type
1 88.5 11.5 0 88.5 remote
2 11.5 77.8 5.1 77.8 ""
3 94.4 0 5.6 94.4 ""
4 23.1 74.1 2.8 74.1 ""
end
list

drop max inst_*

egen max = rowmax(in_person remote hybrid)
gen inst_type = ""
replace inst_typ = "in_person" if max == in_person
replace inst_typ = "hybrid" if max == hybrid
replace inst_typ = "remote" if max == remote

list, sep(0)

And I was not able to replicate your errors. This is what I got:

+-------------------------------------------------------------+
| distri~d remote~t hybrid~t in_per~t max inst_t~e |
|-------------------------------------------------------------|
1. | 1 88.5 11.5 0 88.5 remote |
2. | 2 11.5 77.8 5.1 77.8 hybrid |
3. | 3 94.4 0 5.6 94.4 remote |
4. | 4 23.1 74.1 2.8 74.1 hybrid |
+-------------------------------------------------------------+

Consider using -dataex- to post some sample data that can let us reproduce that error.

Thanks for the help! I was able to solve the issue with Ali's suggestion.