I think the question here is what we are actually expecting to be able to glean from the data. We do not presume that anyone is using sterile technique for this experiment. On the contrary, we expect microbes from the environment (air, water, cooking utensils,- even bodies!) will factor into the microbes that colonize the starter. We have done more controlled experiments (on a necessarily much smaller scale) where we had bakers make starters using flour from the very same batch of flour (http://robdunnlab.com/…/belgian-sourdough-experiment/), to address some of these issues.  Experimental design is always about tradeoffs. We decided to opt for the design with the lowest barier to entry, so as to have the most amount of data. When you have a little bit of data, a little bit of variation matters a lot. When we have a lot of data, we have more statistical power, and so a little bit of variation is more acceptable. We could certainly try a follow-up experiment in which people keep two starters, both from the same flour, discarding and feeding for the first four days on one, and not disturbing other! Id be very interested to see those results!