Usually, I use Fisher's exact test p value when the sample size is small for Pearson chi-square test. Then I realized that IBM SPSS reports exact p values for all the tests (i.e. Pearson Chi-Square, Likelihood ratio, Fisher's exact test, Linear-by-Linear association).
I especially wonder if anyone knows how IBM SPSS calculates exact p value for Pearson Chi-Square statistic (here p=0.042). I also realized Fisher's Exact Test has a test statistic value (here 6.143), which surprises me, since (as far as I know) this test calculates p value directly from the hypergeometric distribution.
I also read the IBM SPSS Exact Tests manual, but I'm still without a clue (http://www.sussex.ac.uk/its/pdfs/SPSS_Exact_Tests_21.pdf).
Any help would be appreciated.