I was about to say the same thing, since I know a fair number of people who experienced a "crisis of faith" at around 30.

However, the second chart seems to correct for cohort effect, i.e. fewer people are becoming more religious as they age than a decade ago