Command for Listwise deletion

[gi peters]

I am doing linear regression and my IV has more cases than my DV so our instructor told us to do listwise deletion. I searched for the command using

help listwise

but it says nothing found. How can I do listwise deletion?

[Clyde Schechter]

I don't understand why your instructor would tell you that. When you run any estimation command in Stata, listwise exclusion for missing values of an involved variables is done automatically. You don't have to explicitly do anything to make that happen.

[Bruce Weaver]

To demonstrate what Clyde explained in #2, try this:

* Generate some data to illustrate
clear *
matrix C = (1, .3 \ .3, 1)
corr2data x y, n(100) corr(C)
* Change a few y-values to missing to mimic the situation described in #1
replace y = . if inlist(_n,7,26,31,77,92)
regress y x
summarize x y
* Use e(sample) to list only the estimation sample from -regress-
summarize x y if e(sample)

Output from the two -summarize- commands:

. summarize x y

    Variable |        Obs        Mean    Std. Dev.       Min        Max
-------------+---------------------------------------------------------
           x |        100    1.83e-09           1  -2.332422   2.238905
           y |         95   -.0300106    .9925115  -2.366965   2.713493

. summarize x y if e(sample)

    Variable |        Obs        Mean    Std. Dev.       Min        Max
-------------+---------------------------------------------------------
           x |         95   -.0119315    1.007608  -2.332422   2.238905
           y |         95   -.0300106    .9925115  -2.366965   2.713493

[Bruce Weaver]

Adding to the example in #3, try this:

regress y x
regress y x if !missing(x,y)

[gi peters]

Bruce Weaver what does the ! indicate in !missing?

[Nick Cox]

! is a logical operator meaning "not".

help operators

[Joro Kolev]

And if you want to do listwise deletion manually for analysis that involves three variables y, x, z, you do

drop if missing(y,x,z)