There are various forms of reanalysis. I am adopting ideas from Hamermesh (2007), Clements (2017), and Christensen and Miguel (2018) to suggest a basic taxonomy: (1) Reanalyses may or may not use the same computer logic (code) as the original paper. (2) Reanalyses may or may not use the same sample, and may or may not draw from the exact same population. These two dimensions suggest at least four categories: reproductions, extensions (including updates), replications, and reexaminations.