Lesson Plan

Compute a difference in summary statistics (means or proportions) between two treatment groups in a randomized experiment

State the no-effect assumption -- that any observed difference is due only to the random assignment -- as the model to be tested

Build a re-randomization distribution by repeatedly shuffling group labels and recomputing the difference under the no-effect assumption

Locate the observed difference in the re-randomization distribution and decide whether it is significant (unlikely to arise by chance alone)

Interpret a significant or non-significant result causally, given random assignment, and state the limits of the conclusion

HSS.IC.B.5

HSS.IC.B.5: Use data from a randomized experiment to compare two treatments; use simulations to decide if differences between parameters are significant.