Answer
The design was flawed.
Work Step by Step
The problem with this experiment is that, while there may be 6 degrees of separation, there is no guarantee anyone will know which people to choose to reach the targeted person. In addition, he should have used a more random sample that included a larger part of the US population.
