How to separate data from one variable into two, without missing in STATA?

[Emma Rohwedder]

I have a dataset with answers form 2004 and 2016 in the same variable, but I want to separate the answers by year without missing values in my two new variables
… Read more
Dataset

[FernandoRios]

Hi Emma
Whatever you were trying to do, is not properly linked.
So why don't you explain what is the problem you are facing. so one may provide you with some advice.
Fernando

[Emma Rohwedder]

Hey again Fernando

I simply want to separate my gender variable into two different ones for each year like this (see picture)

But if I use the separate command it codes the other year into missing values (see picture two)

Is it possible to run the separate command, where it creates two sets of variables without including the opposite year in each variable as missing values?

I hope this explains my problem

[FernandoRios]

I think i understand the problem. The solution, however depends on what is what you want to do.
RIght now your data set organized as follows:

gender   year  
1         1
2         1
1         2
2         2

Using separate, this is what you get:

gender   year  gender1   gender2
1         1      1         .
2         1      2         .
1         2      .         1
2         2      .         2

So you could replace the missing values with anything else, doing something like:

recode gender1 gender2 (.=0)

But something tells me that isn't what you want. what else can you tell us about the task you are trying to do?
Fernando

[Emma Rohwedder]

Thank you for your help. It wasn't quite what I'm looking for.

I have a cumulative dataset, and wish to separate the observations from 2004 and 2016 independently (hence separate the two). When I do the separate commando on one year, the opposite year's observations are coded to missing value, involuntarily on my part. How do one keep the observations from 2004 and 2016, when separating the two?

Example:
We wish to keep the observations on the gender variable on two different years.

	CURRENT VARIABLE		NEW VARIABLE 2004		NEW VARIABLE 2016
2004	MALE		2004	MALE	2016	FEMALE
2004	MALE		2004	FEMALE	2016	FEMALE
2004	FEMALE		2004	FEMALE	2016	MALE
2004	MALE		2004	MALE	2016	MALE
2016	MALE		2004	MALE	2016	MALE
2016	FEMALE		2004	MALE	2016	MALE
2016	MALE		2004	MALE	2016	FEMALE

I wish to keep both observations in 2004 and 2016, as illustrated in the table above. I hope this explains it further.

Thanks in advance!

[FernandoRios]

the problem is not very clear yet.
How do you know that observation 1 will have the value 2016female or observation 5 will have 2004male?

[Emma Rohwedder]

Hi again Fernando

I have attached a picture of the browse window of my separated variables.

I only want the Gender2004 variable to include observations for 2004 and not code the 2016 observations into missing values
Likewise for Gender2016 variable, only to include the data for 2016 and not code the 2004 observations into missing values

Any idea how I separate Gender into the two variables (Gender2004, Gender2016) without getting the missing values for the opposite year included in the new variables?

My commands have been:
keep Gender Year_Of_Study
keep if Gender == 2004 | Gender == 2016
separate Gender, by(Year_Of_Study) // generates the two new variables (Gender2004, Gender2016)

Thank you very much for helping me out!