t-test is equal for paired and unpaired samples

[MawiraIke]

The Welch's test is used to calculate the t-test for paired samples in incanter.stats.

(let [t-stat (if one-sample?
                 (/ (- x-mean mu) (/ (sqrt x-var) (sqrt n1)))
                 ;; calculate Welch's t test
                 (/ (- x-mean y-mean) (sqrt (+ (/ x-var n1) (/ y-var n2)))))

A dependent t-test for paired samples can be used since the samples are dependent. The Welch's test then can be used for independent samples.
Fixing this on my fork and making a PR.

[MawiraIke]

While fixing this issue i made the observation that var-equal option in (t test ) function in incanter.stats is not used. Perhaps it was to be used to determine samples with equal variances and then use a Student T-test and not Welch's test? The current operation is all calculations are done using the Welch's test unless it is a one sample.