As reproductive rights and abortion access again face significant challenge from the politicized and asymmetrical Supreme Court — who decided not to block the enforcement of the Texas “bounty” abortion law — new analysis finds there is not a single state in the nation where support for a federal ban on abortion has more than 30% support among the public.
Data for Progress used raw data from the 2020 Congressional Election Study to estimate state-level support for a federal ban on abortion. The estimation was done using a multilevel-regression model (MRP).
MRP consists of two steps. First, we fit a multilevel-regression model to a survey sample to estimate how likely someone is to hold a particular opinion based their self-reported demographic characteristics and the state where they live. A regression model is a statistical model used to analyze the relationships between some observed outcome (in this case, a political opinion) and other characteristics, called predictors. Multilevel regression is an advanced modeling technique that makes efficient use of sample data.
The second step in MRP is called post-stratification, where we use our multilevel model to predict the responses of the population of each state. Using the demographics and population counts from the voter file, which account for differences in the demographics and geography, we post-stratify the survey sample and the population.
Written by Johannes Fischer and Ahmad Ali. Cross-posted from Data for Progress.