Nonparametric statistical tests, 1 or 2 samples#

Sign test: p-value and confidence interval#

ctx.sign_test(x, k, n, method='default')#

where ctx is dec, mpm, or gmp.

Returns the results of the Sign test, under \(H_0\). See also: Wikipedia [1583], Wikipedia [1566].

Sign test: power and sample size#

ctx.sign_test_power(x, k, n, method='default')#

where ctx is dec, mpm, or gmp.

Returns the results of the Sign test, under general alternatives. See also: Wikipedia [1583], Wikipedia [1566].

Brown-Mood median test for 2 independent samples: p-value and confidence interval#

ctx.brown_mood_test(x, k, n, method='default')#

where ctx is dec, mpm, or gmp.

Returns the results of the Brown-Mood median test for 2 independent samples, under \(H_0\). See also: Wikipedia [1576].

Brown-Mood median test for 2 independent samples: power and sample size#

ctx.brown_mood_power(x, k, n, method='default')#

where ctx is dec, mpm, or gmp.

Returns the results of the Brown-Mood median test for 2 independent samples, under general alternatives. See also: Wikipedia [1576].

Wilcoxon’s signed rank \(T\) test: p-value and confidence interval, continuous data#

ctx.signed_rank_test(x, k, n, method='default')#

where ctx is dec, mpm, or gmp.

Returns the results of the Wilcoxon’s signed rank \(T\) test, under \(H_0\). See also: Wikipedia [1585].

We consider \(N\) continuously distributed random variables \(D_i,i=1\ldots N\), with common pdf \(h_0\). In a sample \((d_1,\ldots,d_N)\) of size \(N\) let \(r_i\) be the rank of \(d_i\) in the ordered sample.

The test criterion of Wilcoxon’s Signed Rank is \(T_N=\sum_{i=1}^N S(d_i)r_i\), where \(S(d_i)=1\) for \(x>0\) and \(S(d_i)=0\) for \(x<0\). \(T_N\) can assume values between 0 and \(\tfrac{1}{2}N(N+1)\) in steps of 1.

Wilcoxon’s signed rank test \(T\) (Bennett alternatives): power and sample size#

where ctx is dec, mpm, or gmp.

ctx.signed_rank_power(x, k, n, method='default')#

Returns the results of the Wilcoxon’s signed rank test \(T\), under Bennett alternatives. See also: Wikipedia [1585].

Mann-Whitney \(U\) test (stratified): p-value and confidence interval, continuous data#

ctx.mannwhitney_test(x, k, n, method='default')#

where ctx is dec, mpm, or gmp.

Returns the results of the Mann-Whitney \(U\) test, under \(H_0\). See also: Wikipedia [1575].

See also: Mehrotra (2006): Rank-Based Analyses of Stratified Experiments: Alternatives to the van Elteren Test

Let \(x1,\ldots,x_m\) and \(y1,\ldots,y_n\) be two sets of measurements, which we denote by \(X\) and \(Y\). The test criterion \(U\) of the Mann-Whitney test is then

\[U = \sum_{i=1}^m \sum_{j=1}^n \text{sgn}(x_i - y_j)\]

Mann-Whitney \(U\) test: (Lehmann alternatives): power and sample size#

ctx.mannwhitney_power_lehmann(x, k, n, method='default')#

where ctx is dec, mpm, or gmp.

Returns the results of the Mann-Whitney \(U\) test, under Lehmann alternatives. See also: Wikipedia [1575].

Mann-Whitney \(U\) test: (Milton alternatives): power and sample size#

ctx.mannwhitney_power_milton(x, k, n, method='default')#

where ctx is dec, mpm, or gmp.

Returns the results of the Mann-Whitney \(U\) test, under Milton alternatives. See also: Wikipedia [1575].

Siegel-Tukey test: p-value and confidence interval#

ctx.siegel_tukey_test(x, k, n, method='default')#

where ctx is dec, mpm, or gmp.

Returns the results of the Siegel-Tukey test, under \(H_0\). See also: Wikipedia [1582].

See also: (Lowenstein 2017)

Kendall test for 2 correlated samples: p-value#

ctx.kendall_test(x, k, n, method='default')#

where ctx is dec, mpm, or gmp.

Returns the results of the Kendall test for 2 correlated samples, under under \(H_0\). See also Wikipedia [1571].

Let \((X_1, Y_1),...,(X_N, Y_N)\) be independent random variables, the \(X_i\) with a continuous distribution \(F_0\), the \(Y_i\) with df \(G_0\). Let \(R_i\) and \(S_i\) be the ranks of \(X_i\) and \(Y_i\), respectively. The Kendall rank correlation coefficient \(\tau\) is defined as

\[\tau = \frac{1}{N(N-1)} \sum_{i=1}^N \sum_{j=1}^N \text{sgn}(R_i - R_j) \text{sgn}(S_i - S_j)\]

and its transformation \(T_N\) is defined by \(T_N = \tfrac{1}{4} (\tau+1)N(N-1)\). \(T_N\) can assume values between 0 and \(N(N-1)/2\).

Theill test for 2 correlated samples: p-value and confidence interals#

ctx.theill_test(x, k, n, method='default')#

where ctx is dec, mpm, or gmp.

Returns the results of the Theill test for 2 correlated samples, under under \(H_0\).

See also: https://en.wikipedia.org/wiki/Theil%E2%80%93Sen_estimator