The digitized Revenue and Expenses Datasets are the only source of data on the break-down of total contributions (p1tcont) into direct public support (p1dirsup), indirect public support (p1indsup), and government grants (p1govgt).
However, the researcher must use these variables with caution, because all Form 990-EZ filers included in the digitized Revenue and Expenses datasets do not report direct and indirect support and government grants, but only report total contributions (p1tcont). For all Form 990-EZ filers (where formtype='Z') the p1dirsup, p1indsup, and p1govgt variables are filled with zeroes, because the actual Form 990-EZ does not have line items 1a, 1b, and 1c for the break-down of contributions. The Form 990-EZ only has Line Item 1= 'Contributions, gifts, grants, and similar amounts received."
Suggestions for assigning a value to p1dirsup, p1indsup, and p1govgt for all EZ filers: If you need to estimate the private contributions (p1dirsup + p1indsup) and government grants received by small organizations by major NTEE category (or by any other control variable), you should take the ratios of (p1dirsup + p1indup)/p1tcont and p1govgt/p1tcont for all Form 990 filers (where formtype <> 'Z') with total expenses <= $100,000 (or any other measure for "small" organizations). Then, apply the ratios to total contributions of EZ filers (where formtype = 'Z') by NTEE category to get an estimate of private contributions and government grants received by small organizations.
NOTE: The same issue is true for the break-down of total contributions into cash (p1cash) and non-cash/in-kind donations (p1noncsh). The digitized Revenue and Expenses datasets show p1cash=0 and p1noncsh=0 for all EZ filers in the data because the Form 990-EZ does not have these line items. The same method of applying the ratios p1cash/p1tcont and p1noncsh/p1tcont for small F990 filers (e.g. exps<= $100,000 and formtype <> 'Z') to the total contributions reported by F990-EZ filers can be used to estimate cash and in-kind donations for small organizations.