I'm talking about the more recent (and more comprehensive and granular) data from NPORS:

"In April, for example, Pew Research Center released data showing a continuation of the familiar pattern in party identification — more liberal to more conservative as the age of respondents increased.

Earlier this week, though, Pew released the most recent iteration of its big, comprehensive National Public Opinion Reference Survey. The NPORS is a benchmark poll that uses a number of different methodologies to ensure responses, including phone, online contact and direct mail. And, as The Washington Post’s Lenny Bronner found after parsing the data, it found a much more complicated interplay between partisan identification and age."

You can read the numerous articles on the data from NYT or the Washington Post if you'd like.