Encoding of the expenses/revenue data

[Johannesderuig]

Not sure if this is the right place for this "issue", but nonetheless here we go:

I am have downloaded some of the e-filer datasets and am currently in the process of putting them together and cleaning them. I noticed that a lot of the (financial) variables are automatically put into strings due to a lot of missing values noted as "NA".
Especially within the expenses and revenue datasets I am not sure how to deal with these "NA" observation.
It seems to me that sometimes these are, as to be expected, just missing values and can be treated as such.
In other cases however, it seems they are a way of denoting 0.
How can/should I distinguish between the "true" NA's and the NA's that are actually just 0? Or are they all just 0?
Any help would be greatly appreciated.

Kind regards,
Johannes de Ruig.

NB: I am not much of a coding expert and work in Stata

[Thiyaghessan]

Hi Johannes, that is a good question. The "NA" observations denote fields left blank by the filer. In some cases, this is the filer's way of expressing a 0; in other cases, the filer did not input a value. We report the inputs as is, with NAs representing blanks. You'll have to look at the descriptions of specific fields to determine if these can be recoded as 0.

[lecy]

Also look at the field scope in the data dictionary. If listed as PZ that means the field exists on both forms. If listed as PC it only exists on the full 990, so any data coming from 990EZ fields would be NA for “not applicable” instead of “not answered”. Only impute zeros for 990 filers in those cases.