mismatch between documentation and code in thinkstats2.Jitter
rjoberon opened this issue · 1 comments
rjoberon commented
The documentation for Jitter states Jitters the values by adding a uniform variate in (-jitter, jitter) but the code does
np.random.normal(0, jitter, n) + values
which is normal distributed data which is not necessarily (rather unlikely) in the range (-jitter, jitter).
Either the documentation should be changed or the code to
np.random.uniform(-jitter, jitter, n)
(which would also better fit the purpose in Ch. 5 to "undo" rounding)
AllenDowney commented
Fixed. Thank you!