The Mitofsky report looking at seven different polling firms' results suggests that we should be pretty skeptical of Reforma's ground-breaking survey from last week. Of course, it has the smallest gap between Peña Nieto and AMLO, at four points, when the average difference was 16 points, and the next smallest was 12 points, from UnoTV. But it also shows the largest swings from March to May, for all of the candidates. Peña Nieto dropped seven points according to Reforma, while the next largest drop was a mere 3.1 percentage points, and the average was just 2.8. AMLO leaped by 12 points with Reforma, while no one else gave him a bump of more than six. Reforma found a drop of 9 points for Vázquez Mota, compared to an average of 5.3 points, and a second-largest drop of six points. Even Quadri, largely an afterthought, had a larger jump than any other candidate.
In short, these are big differences across the board. I'm no Nate Silver, but to me that suggests there was something funky in their sample.