Deleting Observations

[Prince Atorkey]

Hi Everyone,
Please what code should I use if I want to have only observations where participants responded "Yes" to suicidal ideation, "Yes" to suicidal plan and then indicated the number of times they attempted suicide? I tried to use the drop if command, but it deleted all the participants who responded "No" and I cannot do a regression analysis if all those with "No" are deleted.

Ideation Plan Number of time attempted.
1531. Yes No 0 times
1532. No Yes 2 or 3 times
1533. No No 0 times
1534. Yes Yes 1 time
1535. No No 0 times

In the table above, how do delete observations/participants 1531, 1532, 1533, and 1535? I hope this clarifies things.

Thank you for your help in advance.

[Nick Cox]

Data example please using dataex -- the answer depends crucially on whether these variables are string or numeric with value labels.

See https://www.statalist.org/forums/help#stata

[Prince Atorkey]

Thanks, Nick for your response. I am not sure about how to use the dataex command. The variables are numeric with value labels. Thanks

[Nick Cox]

help dataex

That's not precise enough for me to suggest precise code.

[Prince Atorkey]

Hi Nick, I used dataex and generated this. I hope it clarifies my question. Thank you

* Example generated by -dataex-. For more info, type help dataex
clear
input byte(Q24 Q25 Q26)
2 2 1
1 2 4
. 2 3
2 2 2
2 2 1
. 1 4
2 . 1
2 2 1
2 2 1
. . 1
2 2 1
. . 1
2 2 1
2 2 1
2 2 1
2 2 1
1 2 1
2 2 1
2 2 1
2 1 1
2 2 1
2 2 1
2 2 1
2 2 1
2 1 1
2 2 1
2 2 1
2 2 1
2 2 1
1 1 2
2 2 1
2 2 1
2 2 1
2 2 1
2 . .
2 2 1
2 2 1
2 2 1
2 2 1
2 2 1
2 2 1
2 1 1
2 2 1
2 2 2
2 2 1
2 1 1
2 1 1
2 2 1
2 2 1
2 2 1
2 2 1
2 2 1
2 2 1
2 2 1
2 2 1
2 2 1
2 2 1
1 1 1
2 2 1
2 2 1
2 2 1
2 2 1
2 2 1
2 2 1
2 2 1
2 2 1
2 2 1
2 2 1
2 2 1
2 2 1
2 2 1
2 1 2
1 1 3
2 2 1
2 2 1
2 2 1
1 1 1
1 1 2
2 2 1
2 2 1
2 2 1
2 2 1
2 2 .
1 1 1
1 1 3
2 2 1
1 2 1
1 1 5
2 2 1
2 2 1
2 2 1
2 . 2
2 2 1
1 1 2
2 2 1
2 2 1
2 2 1
2 2 1
1 1 3
2 2 1
end
label values Q24 Q24
label def Q24 1 "Yes", modify
label def Q24 2 "No", modify
label values Q25 Q25
label def Q25 1 "Yes", modify
label def Q25 2 "No", modify
label values Q26 Q26
label def Q26 1 "0 times", modify
label def Q26 2 "1 time", modify
label def Q26 3 "2 or 3 times", modify
label def Q26 4 "4 or 5 times", modify
label def Q26 5 "6 or more times", modify

[Nick Cox]

Thanks for the detail.

keep if Q24 == 1 & Q25 == 1

seems to be what you're asking for. Missings on Q26 will be ignored any way in model fitting. But why do that? Just call up your model fitting command with

....  if Q24 == 1 & Q25 == 1

[Prince Atorkey]

Thanks Nick. The idea is to have participants who responded to Yes to Q24 and Q24 and went on to respond to Q26 by indicating the number of times. Then do a regression to see what factors are associated. At the moment, when I use the command suggested, it deletes all "No" responses and regression can't be done. See below the output when i attempted doing regression.

. regress Q24 i.Q2 i.Q1

Source | SS df MS Number of obs = 270
-------------+---------------------------------- F(8, 261) = .
Model | 0 8 0 Prob > F = .
Residual | 0 261 0 R-squared = .
-------------+---------------------------------- Adj R-squared = .
Total | 0 269 0 Root MSE = 0

----------------------------------------------------------------------------------------
Q24 | Coefficient Std. err. t P>|t| [95% conf. interval]
-----------------------+----------------------------------------------------------------
Q2 |
Female | 0 (omitted)
|
Q1 |
12 years old | 0 (omitted)
13 years old | 0 (omitted)
14 years old | 0 (omitted)
15 years old | 0 (omitted)
16 years old | 0 (omitted)
17 years old | 0 (omitted)
18 years old or older | 0 (omitted)
|
_cons | 1 . . . . .
----------------------------------------------------------------------------------------

[Nick Cox]

Indeed. I thought you had other covariates but once you reduce any covariate to a constant it can't appear in a regression. You need to use all the relevant data.

[Prince Atorkey]

Thanks for your assistance Nick.