I expected that the Hispanic share of long-settled families would explain why the share of US adults raised Catholic rose from 2000 to 2024. The General Social Survey does not support that. From 2000-09 to 2021-24, the third-plus-generation non-Hispanic white raised-Catholic rate adds +1.46 points [+0.38, +2.48]. The third-plus Hispanic rate term is -0.30 [-0.57, -0.03], because their raised-Catholic share fell from 68% to 59%. The foreign-born rate term is -1.78 [-2.37, -1.08].
The Hispanic effect that remains runs through composition (4.3% of adults in 2021-24, up from 3.0%), which this split does not separate. The third-plus Hispanic samples are small (367 to 462 respondents), the religion at 16 is recalled, and there is no fertility measure yet, so the question of fertility versus immigration is still open. Code, tables and the log are in research/catholic-demography/ (experiment m5d_ethnicity).