Keep the value within same ID

[HS JANG]

Hello Statalist,

I I've tried many things and I do not know what to do and ask for help now.

My datasets are like this:

ID	Year	treatment	Var3
1	4	0	...
1	5	1	...
2	7	0	...
2	8	1	...

And I would like to classify the treatment with 1 as the group among the IDs whose year is less than 5 years.

What I mean is that, keep ID 1 and drop ID 2, because treatment year of ID 1 is 5, and ID 2 is 7.

ID	Year	treatment	Var3
1	4	0	...
1	5	1	...

Could you please give me some hints or inspirations


Thanks in advance.

[Carlo Lazzaro]

Your query seems (to me, at least) a bit inconsistent, as the longest treeatment spell for ID 1 is 5 years (not < than 5 years, as per your previous statement).
Hence, you should be more detailed about what you're after and, please, di use -datex- to post example/excerpt of your data. Thanks.

[HS JANG]

Hello Carlo,
Thank you for helping me.
My datasets are kind of panel data, and I want to keep the datagroup, which includes treatment is 1 & period <=5.

ID	period	treatment
34063	22	1
39725	23	1
39725	24	1
39725	25	0
39725	26	0
39725	27	0
39725	28	0
39880	1	0
40811	2	0
40811	3	1
40811	4	1
40811	5	1
40811	6	1
40811	7	0
40811	8	0
41315	1	1
41315	2	0
41357	11	0
41357	12	0
41357	13	0
41357	14	0
41357	15	0
41357	16	0
41522	11	0
41522	12	1
41522	13	0
41522	14	0
41522	15	0
41522	16	0
41522	19	0
42000	3	0
42000	4	0
42000	5	1
42000	6	0
43494	7	0
43494	22	0
43494	23	0
43494	24	0
43494	25	1
43494	26	1

Therefore my dataset will be as below:
Groups including treatment before 5 years are only alive.

ID	period	treatment
39880	1	0
40811	2	0
40811	3	1
40811	4	1
40811	5	1
40811	6	1
40811	7	0
40811	8	0
41315	1	1
41315	2	0
42000	3	0
42000	4	0
42000	5	1
42000	6	0

[HS JANG]

Thanks in advance. I'd like to use fixed-effects analysis(xtreg) in group with treatment years less than 5 years.

[Carlo Lazzaro]

There's still something I'm not clear with in your data.
For instance: id 39880 matches the time criterium (period<=5) but not the treatment one (0 vs 1) (your criteria were: treatment== 1 & period <=5).
Please, clarify.

[HS JANG]

Hi Carlo,

I made some mistakes describing my data.
My criteria were treatment==1 & period<=5. So the ID 39880 should be removed.

Here are my data that I want to make:

ID	period	treatment
40811	2	0
40811	3	1
40811	4	1
40811	5	1
40811	6	1
40811	7	0
40811	8	0
41315	1	1
41315	2	0
42000	3	0
42000	4	0
42000	5	1
42000	6	0

[Carlo Lazzaro]

Despite your laudable efforts to make things clearer, you still have weird results in your desired database.
Be as it may, according to your criteria you should end up with what follows:

. egen flag=group( period treatment) if period<=5 & treatment==1


. drop if flag==.


. list

     +----------------------------------+
     |    ID   period   treatm~t   flag |
     |----------------------------------|
  1. | 40811        3          1      2 |
  2. | 40811        4          1      3 |
  3. | 40811        5          1      4 |
  4. | 41315        1          1      1 |
  5. | 42000        5          1      4 |
     +----------------------------------+

[HS JANG]

Dear Carlo,
It's fantastic! I cannot find that in any books.
Thank you very much!

[Carlo Lazzaro]

You're welcom.
It's only a matter of practicing!

[Carlo Lazzaro]

Previous welcom is welcome.
Sorry for the typo.