Well, Microsoft's test completely backfired on me. I ran the test many times. In most cases, one of the columns had clearly better results, and 80+% of the time, that column was Google.
Small sample size to be sure, but it covers 100% of the population I care about in choosing a search engine (me).
Granted, this was Bing, not Google, but I kind of doubt DDG would fare better. (And in non-blind tests, I definitely fell I'm getting the worst results from DDG).