fisher_exact

mpsci.stats.fisher_exact(table, alternative='two-sided')

Compute Fisher’s exact test on the 2x2 table.

This function uses python integers to compute the probability exactly. (Unlike most of the other functions in the mpsci library, this function does not use mpmath.)

Returns the sample odds ratio and the p-value. Generally the return values will be instances of fractions.Fraction, but the odds ratio will be math.nan if both a*d and b*c are zero, and it will be math.inf if b*c is 0 and a*d is not zero, where a = table[0][0], b = table[0][1], c = table[1][0] and d = table[1][1].

Warning: The values in table should not be large! The calculation can take a very long time and use a lot of memory if the values are large. (The precise definition of “large” will depend on the computer’s speed and available memory.)

Examples

>>> from mpsci.stats import fisher_exact

>>> table = [[8, 2], [1, 5]]
>>> oddsratio, pvalue = fisher_exact(table)
>>> oddsratio, float(oddsratio)
(Fraction(20, 1), 20.0)
>>> pvalue, float(pvalue)
(Fraction(5, 143), 0.03496503496503497)

An example with bigger values:

>>> table = [[345, 455], [260, 345]]
>>> oddsratio, pvalue = fisher_exact(table)
>>> pvalue  # The following output is edited for this docstring.
Fraction(5815594848535879587533843493536161337167965268243476365254048617
5074871119589194456724424702642108585451894970002396727579577842267797518
4929838662845360913926006683618646889965042763561174204677207235122556332
9503454183984380902895184450793815543305546132593168918382926520280514271
135809969597, 60789478542597481225083508544099336316985118448918449312948
2297975307374215835510960358633704364890832917577706721466773385015807270
0232888943268164264596951845578381273218929422021631512908332122679219965
7632047453216230717808020513225358000202208441824610229474920736533616413
69838473616509597)
>>> float(pvalue)
0.9566778639926435