statsmodels.stats.power.TTestIndPower.power#

TTestIndPower.power(effect_size, nobs1, alpha, ratio=1, df=None, alternative='two-sided')[source]#

Calculate the power of a t-test for two independent sample

Parameters:

effect_sizefloat: standardized effect size, difference between the two means divided by the standard deviation. effect_size has to be positive.
nobs1int or float: number of observations of sample 1. The number of observations of sample two is ratio times the size of sample 1, i.e., nobs2 = nobs1 * ratio
alphafloat in interval (0,1): significance level, e.g., 0.05, is the probability of a type I error, that is wrong rejections if the Null Hypothesis is true.
ratiofloat or array_like, optional: ratio of the number of observations in sample 2 relative to sample 1. see description of nobs1. ratio must be non-negative and not NaN. ratio can be set to zero in order to get the power of the one sample t-test with nobs1 observations, in which case the default degrees of freedom are df = nobs1 - 1. This is a convention and not the limit of ratio going to zero: a ratio close to zero, e.g. 1e-9, means that sample 2 has almost no observations and the power is close to alpha, while ratio=0 returns the one sample power.
dfint or float, optional: degrees of freedom. By default this is None, and the df from the ttest with pooled variance is used, df = (nobs1 - 1 + nobs2 - 1) or df = nobs1 - 1 if ratio=0.
alternative{‘two-sided’, ‘larger’, ‘smaller’}, optional: extra argument to choose whether the power is calculated for a two-sided (default) or one sided test. The one-sided test can be either ‘larger’, ‘smaller’.

Returns:

powerfloat or ndarray: Power of the test, e.g., 0.8, is one minus the probability of a type II error. Power is the probability that the test correctly rejects the Null Hypothesis if the Alternative Hypothesis is true. The result is a float for scalar arguments and an array if ratio or another argument is array_like.

Raises:

ValueError: If any element of ratio is negative or NaN.

Notes

If ratio is an array_like, then the one sample convention for ratio=0 is applied to the elements that are zero.

Changed in version 0.15.1: ratio=0 returns the power of the one sample t-test, it used to raise a ZeroDivisionError or to return alpha with a RuntimeWarning. A negative or NaN ratio raises a ValueError.