The other is that they haven't tracked people over time, just made some measurements and extrapolated from that. This would seem to suggest that the conclusion should be more like "Older people more liberal nowadays". Or as John Allen Paulos points out do we believe the statistics that tell us:
since half the people in the U.S. are men and half are women, the average American adult has one ovary and one testicle, or the average resident of Dade County, Florida is born Hispanic and dies Jewish.
It does assert in the article that they've corrected for this type of error, but it just seems simpler to conclude that older people today, who were in their teens and twenties in the sixties, are more liberal than the older people of the sixties.