@oristarA,
The math used in the statistics is not available to us, but I can readily see a number of pitfalls. The mortality of any cadre of the population is strongly influenced by its average or median age. What is the average of the reference population "with Degrees"? What is the average age of "everyone over 30?
Even a small difference here could grossly distort the implied mortality difference.
It sounds elementary but errors like this are VERY common in the statistical analyses that make their way into the public domain -- particularly such social analyses produced by folks with an axe to grind or publication stats to accumulate.
Some social "scientists", even with degrees, are quite stupid.
The information age has certainly multiplied the amount of accessible data out there, but the amount of meaningful information in it amounts to a rather large degrements. Worse, it can be hard to separate the wheat from the chaff.