Summary of the Chapter#
In this chapter, we learned how data science uses experiments and observational studies to answer questions. Experiments can show cause and effect because we change something and compare results. Observational studies help us learn about the world without changing anything, but they cannot easily show what causes what.
We also saw that good experimental design must deal with confounders and bias so that comparisons are fair. When experiments are hard to run, surveys and simulations give us other ways to collect information and test ideas. Understanding how studies are designed helps us avoid confusing correlation with causation.