This sounds like you want to use the [McNemar][1] test which is intended for paired nominal (boolean) data. There seems to be a Python implementation in the [statsmodel][2] library. See also [this][3] question. [1]: https://en.wikipedia.org/wiki/McNemar%27s_test [2]: https://www.statsmodels.org/devel/generated/statsmodels.sandbox.stats.runs.mcnemar.html [3]: https://stackoverflow.com/questions/41539164/mcnemars-test-in-python-and-comparison-of-classification-machine-learning-model