Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: more bizarre missing observations

From   David Kantor <[email protected]>
To   [email protected]
Subject   Re: st: more bizarre missing observations
Date   Thu, 09 Oct 2003 09:08:57 -0400

At 12:43 PM 10/9/2003 +0100, Kristien Verheyen wrote:
I have a similar problem to Robert's in that I know for sure there are
missing data for a particular variable ('canter') but if I type

tab canter if canter==.

I am told there are 'no observations'

canter is stored as float but following Neil's suggestion still gives me 'no
observations'. Please can anyone tell me how I find out exactly how many
missing values there are?
Thank you,

Kristien Verheyen
By default, -tabulate- omits missing observations, so the behavior you see is completely expected.

If you want to see the missings, then code...

tab canter if canter==. , mis

which will allow the missings to be shown.

But note that such a tabulation will have only one row: missing (.). You can get the same information by...
count if canter==.

Another issue to consider is, do you use Stata 8 or an earlier edition? If you have Stata 8, then to test canter==. is to seek a specific missing value: . But there are others: .a, .b, .c, etc.. If you want to get all missing values, then you should test...
or, equivalently,
or, equivalently,
canter <.

Thus, a tabulation such as
tab canter if mi(canter), mis
will show you all the different species of missing values. And
count if mi(canter)
will just give you the total, regardless of the species.

I hope this helps.
-- David

(The missing() function exists at least as far back as Stata 6. The mi() function, as a synonym for missing() came later -- I'm not sure if it is Stata 7 or 8.)

David Kantor
Institute for Policy Studies
Johns Hopkins University
[email protected]

* For searches and help try:

© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index