the back-door criterion identifies: sufficient adjustment sets for causal estimation

Can we trust studies without random experiments?

Imagine you want to know if eating chocolate causes acne. You can't run a controlled experiment for obvious reasons.

Causal models let us use observational data to figure out if there's a real connection between chocolate and acne, without needing to experiment on people.

Example

Researchers look at people who eat chocolate and those who don't, considering other factors like diet and stress, to see if there's a pattern linking chocolate to acne.

Remember this

Causal models help us understand cause-and-effect relationships using real-world data.

Text adapted from Wikipedia, licensed under CC BY-SA 4.0.

Related concepts

the condition number κ(A) measures: sensitivity of Ax=b to perturbations

How small changes affect big outcomes

an instrumental variable does: isolates causal effect when you can't randomize

Why can't we always trust what we see?

Prior probability

Why do we sometimes need to start with an educated guess before diving into new data?

Bootstrapping (statistics)

Small sample sizes can mislead standard error estimates

Causal model

Causal models use DAGs to represent causal relationships

Race and intelligence

IQ test performance differences between racial groups have decreased over time

Swipe through more Machine Learning concepts

Open Pocket Polymath