While the article certainly lacks hard proof, the "he didn't pick the right kids" argument is even worse. The ratio of successful college grads to non-grads may not show causality, but at least it is some sort of metric. Until someone can quantify "the right kids", it beats the alternative.