It’s my statistic, actually, and it’s based on eligible voters, not on registered voters.
After the 2012 election, which Romney lost by 3.9%, I decided to calculate what the popular vote would have been if every eligible voter in the USA had voted.
I used the racial, gender, and age statistics collected by NEP, which is the exit pollster used by all major USA news media.
I used the eligibility stats calculated by Pew Research.
And I used the 2012 population stats from the US Census Bureau.
My result?
Romney lost by 4.1%.
In my personal opinion, that number is low, and is probably 5% or more.
Why?
Historically, white Conservatives turn out at a higher rate than any other ethnic or political group.
To me, that indicates that a 100% turn out would bring more Left wing voters to the polls, and fewer Conservatives.
Did you take into consideration the vast quantity of Bible belt voters, the evangelicals if you will that DID NOT VOTE due Romney being a Mormon into your figures? THAT would make your result moot.