add illustration of approximate vs exact p-vals

README.md

@@ -44,11 +44,13 @@ Ordinal and numerical
 In permutation tests, the null distribution is sampled by random permutations
 of measurements. Making no hypothesis on the statistic distribution, the
 p-value is defined as the proportion of the null distribution with test
-statistic greater (or lower, depending on the test) than or equal to the test
-statistic of the observed data.
+statistic greater (or, for certain test statistic, lower) than or equal to the
+test statistic of the observed data.
 When all permutations are evaluated, these permutation tests give
 [exact tests](https://en.wikipedia.org/wiki/Exact_test).
 
+![](/doc/fig_approx_vs_exact.png)
+
 In case of several variables, they naturally perform a correction for
 [multiple tests](https://en.wikipedia.org/wiki/Multiple_comparisons_problem):
 - rmax permutation test from Pearson r correlations,
@@ -93,7 +95,7 @@ Hall/CRC, 2007.
 
 ## Documentation
 
-TODO: put doc on RTD.
+TODO: put documentation on RTD.
 
 ## Installation
 
@@ -116,5 +118,5 @@ Use `pytest`.
 
 ## Authors
 
-Quentin Barthélemy, an anonymous researcher and Louis Mayaud, 
+Quentin Barthélemy, an anonymous researcher and Louis Mayaud,
 under the wisdom of [Marco Congedo](https://github.com/Marco-Congedo).

examples/compute_exact_ttest_ind.py

@@ -71,7 +71,7 @@
 plt.legend(loc='upper left')
 ax2 = fig.add_subplot(122, title='Exact p-value', xlabel='t', sharey=ax1)
 ax2.axvline(x=t_exact, c='r', label='Real statistic t')
-y, x, _ = ax2.hist(t_dist, 50, density=True, alpha=0.4,
+y, x, _ = ax2.hist(t_dist, 50, density=True, histtype='step',
                    label='Exact t-distribution')
 ax2.fill_between(x[1:], y, where=(x[1:]>=t_exact), facecolor='r', step='pre',
                  label='Exact p-value')