This analysis is a follow-up to @benbbaldwin's post The accumulation of QB hits vs passing efficiency where his conclusion was that:
[For] a given level of hits, there is no evidence that the accumulation of hits makes any difference throughout the course of a game.
As the title suggests, this analysis will take a different approach to measuring the hypothesized effect.
Here is the code to get the passing data we'll be using from nflscrapR-data:
@nflscrapR's Expected Points (EP) is a popular metric among analysts doing public research of play in the NFL. Detailed in the creators' research paper, the metric is derived from a model that was built as a part of a larger system designed to calculate individual wins above replacement values for offensive skill players.
The authors very graciously made public all of their data (nflscrapR-data) and code (nflWAR, nflscrapR-models, nflscrapR) for this project, including the code used to build the EP model. In the init_ep_fg_models.R file of the nflscrapR-models repository, we can see that the following variables are used