statsmodels.stats.weightstats.CompareMeans.tconfint_diff

CompareMeans.tconfint_diff(alpha=0.05, alternative='two-sided', usevar='pooled')

confidence interval for the difference in means

Parameters:

alphafloat

significance level for the confidence interval, coverage is 1-alpha

alternativestr

This specifies the alternative hypothesis for the test that corresponds to the confidence interval. The alternative hypothesis, H1, has to be one of the following :

‘two-sided’: H1: difference in means not equal to value (default) ‘larger’ : H1: difference in means larger than value ‘smaller’ : H1: difference in means smaller than value

usevarstr, ‘pooled’ or ‘unequal’

If pooled, then the standard deviation of the samples is assumed to be the same. If unequal, then Welch ttest with Satterthwait degrees of freedom is used

Returns:

lower, upperfloats

lower and upper limits of the confidence interval

Notes

The result is independent of the user specified ddof.