This is about what one would expect:  a strong but not perfect correlation between PGR rank and later tenure-stream placement.  Unlike some others, Professor Weisberg sensibly looks at  the correlation between PGR rank in 2006 with those who got PhDs six to eight years later, i.e., those who were relying on the 2006 rankings in choosing programs (an example of an earlier effort that wasn't so sensible).