paired t-test?

This is either a “implement such a test” or a “document it better”:

I’m currently looking for the statsmodel equivalent of (from http://statistic-on-air.blogspot.de/2009/07/paired-students-t-test.html)

a = c(12.9, 13.5, 12.8, 15.6, 17.2, 19.2, 12.6, 15.3, 14.4, 11.3)
b = c(12.7, 13.6, 12.0, 15.2, 16.8, 20.0, 12.0, 15.9, 16.0, 11.1)

t.test(a,b, paired=TRUE)

    Paired t-test

data: a and b
t = -0.2133, df = 9, p-value = 0.8358
alternative hypothesis: true difference in means is not equal to 0
95 percent confidence interval:
    -0.5802549 0.4802549
sample estimates:
mean of the differences
    -0.05

I googled for “statsmodels t-test paired” and only found a ttost_paired. Is there simply no paired t-test or am I just not good enough in statistics to find the right test? In the latter case, for the sake of people like me: it would be nice if then the documentation could be adjusted 😃

[josef-pkt]

Yes, as @dengemann pointed out it’s just a one sample t-test on the difference

>>> a = np.array([12.9, 13.5, 12.8, 15.6, 17.2, 19.2, 12.6, 15.3, 14.4, 11.3])
>>> b = np.array([12.7, 13.6, 12.0, 15.2, 16.8, 20.0, 12.0, 15.9, 16.0, 11.1])
>>> stats.ttest_1samp(a - b, popmean=0)
(-0.21330847512739051, 0.83584000571392769)
>>> stats.ttest_rel(a, b)
(-0.21330847512739051, 0.83584000571392769)

>>> from statsmodels.stats import weightstats as sms
>>> d = sms.DescrStatsW(a - b)
>>> d.ttest_mean()
(-0.21330847512739051, 0.83584000571392769, 9.0)
>>> d.tconfint_mean()
(-0.58025487182119562, 0.48025487182119597)

edit stats is scipy.stats, added ttest_rel also

[jankatins]

first of all: thanks for the fast help!

I can’t be the first person to google “paired t-test statsmodels” and felt lost. Especially when I see the results of “paired t-test R”.

It would also be nice if statsmodels would wrap the scipy functions to return a nice printable object which is similar to the regression output (or Rs output) and not only the plain numbers. A function like https://stat.ethz.ch/R-manual/R-devel/library/stats/html/t.test.html would be nice…