You work on a software product where users rely on documentation while writing code. Your team wants to introduce a documentation intervention—such as revised examples, clearer API guidance, or contextual help—and early signals suggest coding quality may improve, but it is unclear whether the documentation change actually caused the improvement.
How would you evaluate whether the documentation intervention caused better coding quality? Walk through the experiment you would run, how you would define success, and how you would avoid drawing the wrong causal conclusion.