Dividing a variable by another one.

[Hatim Tareq]

Hi All,

I have the following table, which is output from the data that I have, I would like to be able to divide the column with name (1) by the total. I'm not sure if I must create a new data set using this output and then do what I need (which I can do that and have no problem), or can I do it in the same date file?

agegrp	0	1	Total
1	144,704	1,266	145,970
2	62,841	980	63,821
3	165,049	5,398	170,447
4	383,228	21,983	405,211
5	330,988	23,963	354,951
6	189,314	16,627	205,941

Regards,

[daniel klein]

This cannot possibly be a dataset, as the names 0 and 1 are not valid Stata variable names. Where does this output come from, i.e. what did you type (exactly) to get this output?

Maybe all you need to do is adding the row option to a tabulate call.

Best
Daniel

[Hatim Tareq]

Hi Daniel,

Sorry for that, the 0 and 1 are the obs of the variable called Died (0 died, 1 alive).

Thanks

[daniel klein]

That does not answer the question of what exactly you typed, and what exactly you want to do here. It is hard to give better advice then.

If you are only interested in the row percentages and you started from

tabulate agegrp died

then, as suggested

tabulate agegr died ,row

will do the trick.

If you have four variables, agegr, died, alive and total, you might be happy with

generate newvar = alive/total

If you situation is still different and you need more help, please explain in more detail.

Best
Daniel

[Hatim Tareq]

Thank you Daniel. What if I want the percentage to six digit (for example: 0.008673)

Regards,

[Svend Juul]

You can use proportion - but this is a rather clumsy solution:

sysuse auto
proportion rep78 , over(foreign) cformat(%9.8f)

[daniel klein]

I still need to assume you have two variables and started with a tabulate call, as you still have not verified this is the case.

There might be implemented ways of doing this easier, but here is one that involves using Mata

sysuse nlsw88 ,clear

tabulate occupation union ,row matcell(X)

m : 
    X = st_matrix("X")
    X = X, (X[., 1] :/ rowsum(X))
    st_matrix("X", X)
end

matlist X

Best
Daniel

[Hatim Tareq]

Hi Daniel, Yes I have two variables (age group and Died).

Thank you for your help, really appreciate it. Is it possible to divided those who are died (1) by the sum of the row (as I get rate of those who are still alive (0) by the sum of the row).

Regards,

[Hatim Tareq]

Thank you Svend for your help. I really appreciate it.

Regards,

[daniel klein]

Is it possible to divided those who are died (1) by the sum of the row (as I get rate of those who are still alive (0) by the sum of the row).

This is what the code I posted does.

Best
Daniel

[Hatim Tareq]

Hi Daniel.

You code divide those who are alive (0) by the sum of the row..

Regards

[daniel klein]

Change

X = X, (X[., 1] :/ rowsum(X))

to

X = X, (X[., 2] :/ rowsum(X))

[Hatim Tareq]

Thank you Daniel,

Is it possible to graph the output (the proportion of those died in the y-axis and the age group in the x-axis).

Regards,