what is assumed (although it also has been tested) is that the random changes in non-selected DNA segments occur at a fairly constant rate. This is because the enzymes involved in DNA copying and repair have a(n extremely small) error rate. This rate has been investigated among and across taxa, and seems for the most part to be very, very consistent. This is actually unsurprising given that those basic enzymatic pathways have been evolved 2-4 billion years ago and are so critically impprtant that selection strongly works against significant changes in them.
As to the article/study, the online article doesn't say much conclusive about what was done specifically, but they do mention the average speciation rate, which indeed seems to be about 2 millions years, regardless of which taxon we're talking about (and this observation is entirely independent from the molecular findings I mentioned above).