FAQ/chiwww - CBU statistics Wiki
location: FAQ / chiwww

How do I obtain Fisher's exact test and chi-square for a two-way table in EXCEL?

Web calculators for 2x2 tables are here and here.

An EXCEL macro may be downloaded from here to add-in to EXCEL for Fisher's exact test for a 2x2 table. The function fet()can then be run in an empty cell to produce the Fisher's p-value. For example fet(A1,B1,C1,D1, True=1-tailed) produces a one-tailed p-value where the four table frequencies are in cells A1, B1, C1 and D1.

Another EXCEL macro written by Russell John which was available from free download from the Murdoch edu website is available below with a slight modification needed to be made to work on CBSU machines. This macro computes Fisher exact tests for any two-way table of frequencies.

In particular the Round () function is not recognized so an amended version is available. This does not run on the CBSU wiki so save this EXCEL file to your PC and run from there. Another amended version here adds an effect size column, esize, which equals Phi (for 2x2 tables) or the contingency coefficient (for other 2-way tables).

The table of frequencies in any two-way table are placed in the EXCEL file and the macro run using tools>macro>macros in older versions of EXCEL (pre-2010) or in EXCEL 2010 by clicking on the developer tab and clicking macros (left hand side). A window showing the TwoWayTable macro should then appear. Choose TwoWayTable Macro and Run. This is also described, with a diagram, on Russell's website (link above). The developer tab is not in the ribbon by default but can be added by clicking the File tab, clicking Options, and then clicking the Customize Ribbon category. In the Main Tabs list, select the Developer check box, and then click OK.

Please note the Monte-Carlo chi-square option in the EXCEL macro does not work on CBSU machines.

A MS-DOS program (fet.exe) which works on all Windows operating systems is available for download or on-line use from here.

None: FAQ/chiwww (last edited 2014-03-19 14:37:00 by PeterWatson)