They've identified the following needs:
- Overall data on how many only-mobile and partly-mobile editors there are. A histogram of editor count by mobile edit proportion is probably the right deliverable.
- Random samples of 200 mobile edits from each of the English, Czech, Korean, French, and Hindi Wikipedias for manual analysis.
- On each wiki, 160 should be mobile web edits, 20 should be Android edits, and 20 should be iOS edits.
- A selection of users Abbey can email for the survey. The population is (1) English Wikipedia editors (2) with email addresses (3) who have made at least 5 edits (4) who have edited in the past 90 days. The stratification by region, platform, and experience is described in this spreadsheet. The data should be: