mqr.inference.rate.

size_2sample#

mqr.inference.rate.size_2sample(r1, r2, alpha, beta, H0_value=None, alternative='two-sided', method='score', compare='diff')#

Calculate sample size for difference of two rates of events.

Null-hypothesis by compare: ‘diff’

r1 - r2 == H0_value

‘ratio’

r1 / r2 == H0_value

Parameters:

r1float: First rate.
r2float: Second rate.
alphafloat: Required significance.
betafloat: Required beta (1 - power).
H0_valuefloat, optional: Null-hypothesis rate. Default 0 for ‘diff’, 1 for ‘ratio’.
alternative{‘two-sided’, ‘less’, ‘greater’}, optional: Sense of alternative hypothesis.
methodstr, optional: ‘z’

Approximation based a normal approximation. For this method, compare must be ‘diff’ and H0_value must be 0.

(other)

All other methods are passed as parameters to power_2sample while that function is solved numerically equal to the requested power by varying nobs.
compare{‘diff’, ‘ratio’}, optional: ‘diff’

Compare the difference between rates.

‘ratio’

Compare the ratio of rates.

Returns: