In keeping with these aims, the top-level or initial survey, hereafter USAGE, consisted of eight questions, whose answers would distinguish different patterns of distribution. Users were asked
Sub groups of respondents were automatically differentiated according to answer-values to one or more questions in USAGE, with these values being treated as single predicates or Boolean combination of predicates. The system dispatched follow up surveys to members of the subgroups which focused on either their downstream (secondary) distribution or their personal use of the documents. Those reporting that they regularly forwarded documents or the information in the documents to 26 or more people were classified as ``distributors.'' If their distribution was online, they received surveys whose questions were keyed to this activity in one of six possible institutional contexts, a predicate which remapped answer values to the institutional affiliation question in USAGE.
The responses to these follow-up instruments were comparable, since many questions were shared across all of them. The questions for these groups concerned
Regardless of institutional affiliations, offline distributors received the same instrument, which added questions about plans of going on line and omitted questions dealing with enhancements.
Respondents to USAGE who distributed to fewer than 26 people were designated ``customers,'' and further classified into two main and two minor groups, each having a corresponding survey instrument:
All four groups were asked for demographic and media-use information, including
The first three groups were also asked about time spent with the documents, ways of selecting them and uses for them in political conversations, work or other activity. The Commercial Customers instrument included questions dealing with rate structures and the Former Customers instrument asked why the user no longer got the documents. As with the Distributor survey instruments, a large percent of questions were shared across all the customer surveys, allowing comparability of responses to them.
A second set of follow up surveys was used to get more finely grained information on distribution. They were sent to USAGE respondents who had indicated a non-MIT-system source for the documents, and were specialized for the type of source indicated and asked for an address of that source. That is, separate forms asked for a Mailing-List address, a Server-Site host-name, a Bulletin-Board telephone number, a Phone-Network address or a Newsgroup name.