6.8.5.2.26. statsmodels.sandbox.stats.multicomp.simultaneous_ci

statsmodels.sandbox.stats.multicomp.simultaneous_ci(q_crit, var, groupnobs, pairindices=None)[source]

Compute simultaneous confidence intervals for comparison of means.

q_crit value is generated from tukey hsd test. Variance is considered across all groups. Returned halfwidths can be thought of as uncertainty intervals around each group mean. They allow for simultaneous comparison of pairwise significance among any pairs (by checking for overlap)

Parameters:

q_crit : float

The Q critical value studentized range statistic from Tukey’s HSD

var : float

The group variance

groupnobs : array-like object

Number of observations contained in each group.

pairindices : tuple of lists, optional

Indices corresponding to the upper triangle of matrix. Computed here if not supplied

Returns:

halfwidths : ndarray

Half the width of each confidence interval for each group given in groupnobs

See also

MultiComparison
statistics class providing significance tests
tukeyhsd
among other things, computes q_crit value

References

[R79]Hochberg, Y., and A. C. Tamhane. Multiple Comparison Procedures. Hoboken, NJ: John Wiley & Sons, 1987.)