There's really no justification for the methodology in the article, which neither: 1. Compares plans with like coverage (so its not comparison of apples to apples in qualitative terms), 2. Compares minimum prices -- instead it compares the median of the five cheapest plans on eHealthInsurance with the median of the lowest tier plans on the exchange (so its not a comparison of the cheapest price for which you can buy anything labelled as "health insurance", even ignoring quality) 3. Considers subsidies available (which is pretty important to the point it is trying to refute, which is that the exchange makes health insurance affordable.)