Fair point - the exact number wasn't very important for my argument (I believe it would still carry even with the 2-4% or 2-4m figure), so I just grabbed the first figure I found. It passed my initial sanity check because I interpreted the "couples" to include "non-married couples", and misremembered the US population to be 500 million rather than 300 million. (~4% of the adult population being swingers didn't sound too unreasonable.)