Perhaps it’s no surprise that polling about America’s religious life is tricky. Religion is personal, nuanced and evolving — all things that make it resistant to easy quantification. But it’s that enigmatic nature of religious life that makes statistical analysis so appealing. It’s easy to be seduced by the promise of reducing faith to one tidy data point: For example, the share of adults who describe themselves as Christians has dropped by nearly 8 percentage points in just seven years. But the polling process is never as tidy as it seems. What does it mean to describe yourself as Christian? How does that self-identification intersect with race, class and community? What does an 8 percentage point drop even mean? The questions that get asked in religion surveys — Do you pray every day? Is religion “old-fashioned”? — are packed with assumptions and terms that don’t mean the same thing to all people.