It would take some detailed study. If there are two people who live together and thus use the same internet access, and if they either use the same computer or the same setup (software versions), checkuser will not be able to distinguish between them. I notice, however, that the blocks do not shut off talk page access, so a wrongfully blocked user could make the point. That Deli nk has not commented is evidence in favor of the sock idea.
They do advise that people in a situation like this disclose it in advance, otherwise checkusers will make the assumption that same access and same user agent string equals same person. The same thing can happen, of course, with library or internet cafe access. If they use the same computer, i.e., they are not editing at the same time, there will be a very similar pattern of editing to being a single person. If they use a different computer, there are still likely to be correlations. To study this, in the past, with other suspected socks, I have loaded the edit times into Excel and plotted them. Truly independent people will show very different patterns from a single person with multiple accounts. It is obvious. By the way, an occasional "edit coincidence" (same minute of edit) does not prove two different people. People may have more than one edit window open. To test it, one time I made a series of edits all in one minute. Not difficult at all. But doing this all the time without showing some odd pattern would be very difficult.
Does anyone really care?