An update on the cookieing front: in my weekend’s worth of data, there were only two 2 visitors—that’s right, only TWO—across the 5000+ total visitors tested who had JS enabled but didn’t have cookies on. It’s definitely possible that there are a decent number of users who have cookies on, but JS disabled, though.

Although my A/A testing is done for the moment, I’m going to make a quick tweak to ShrimpTest to record users who return with a cookie to mark those visitors (in PHP) as having cookie support, even if they don’t have JS on, so I can get a sense of whether that population (+ cookies, – JS) is significant at all. This was a bug in my original methodology… I was only checking for cookie support via JS, so there are many people who probably were [ + cookies, – JS ] but simply marked as [ -cookies, -JS ].