Nick, from my point of view, it is easier to type:
nmissing x
than:
count if x == .
Easier, and nmissing (appropriately updated) gives the right answer,
which count does not...
Note that the count statement as given does not tell you how many
missing values there are; it tells you how many are equal to ".".
There are 27 kinds of missing values in Stata, and if you want to
trap all of them,