Daycia, you're not missing anything - the reference guide does not do a very good job of explaining how to handle this situation. There are a couple of best practices that have been established by capable project teams and GBCI review teams that should be observed:
1) You must sample at least 20% of the fixtures in the building - sorry, there's no way of getting around that one
2) You should test each fixture at least three separate times and use the average flush rate of your three tests for the installed rate in your calculations
3) Depending on the size of your building, you should start your flush rate testing on the ground floor and move up floor-by-floor until you've tested at least 20% of the fixtures. The water pressure in your building will vary with the height of the floor, so it's important to make sure that your sample includes those floors that are expected to have the highest water pressure.
4) Make sure to track all of your testing data in a well-organized spreadsheet and make sure to upload that data, along with a detailed narrative describing your testing methodology, as part of your WEp1 supporting documentation.