RCS Statistician Andrew Marder has compiled some resources for researchers working with observational study data who are interested in identifying causal effects:
Paul Goldsmith-Pinkham has some very nice slides about using difference-in-differences models to identify causal effects: https://paulgp.github.io/presentations/dind.pdf
One paper discussed in the slides, Difference-in-Differences with Variation in Treatment Timing by Andrew Goodman-Bacon (https://www.nber.org/papers/w25018), is particularly interesting. Goodman-Bacon (2018) shows how the causal effect estimated by a model with firm and time fixed effects can be seen as a weighted average of all possible 2x2 difference-in-differences models. For a quick introduction to Goodman-Bacon (2018), see his Twitter thread here!