"It could be that we are under-estimating the mass of the matter that we can see because of an erroneous assumption somewhere earlier in the chain of logic."
I suppose that's not impossible, although it's hard to imagine such an error getting it wrong by a factor of 20, or about 1900% error. Such would be an enormous mistake!
I think the logic is based on trying to reconcile several different measurements of matter content. You can "weigh" the Universe by weighing the galaxies in a number of different ways, and compare that to a prediction based on, say, General Relativity. And they wildly disagree -- but at the same time all of the calculations are tried and tested in other contexts. After so many years, it's hard to see it being just someone forgetting to carry the one.