Automatically perform Fisher's exact test whenever a WARNING message for the Chi-Square test is issued

Contents:

Purpose / Requirements / Usage / Limitations

PURPOSE:

The %RUN_FISHERS macro automatically performs Fisher's exact test when PROC FREQ issues a WARNING message that the Chi-Square test might not be valid due to a large number of cells with small expected frequency.

REQUIREMENTS:

Version 6 or later of Base SAS software.

USAGE:

This macro is only needed when the table that is larger than 2x2 since the CHISQ option in the TABLE statement of PROC FREQ provides both the chi-square test and Fisher's exact test for 2x2 tables.

Follow the instructions in the Downloads tab of this sample to save the %RUN_FISHERS macro definition. Replace the text within quotes in the following statement with the location of the %RUN_FISHERS macro definition file on your system. In your SAS program or in the SAS editor window, specify this statement to define the %RUN_FISHERS macro and make it available for use:

   %inc "<location of your file containing the RUN_FISHERS macro>";

Following this statement, you may call the %RUN_FISHERS macro. See the Results tab for an example.

The following parameters are required:

DATA=

SAS data set that contains the table of interest.

ROW=

The name of the single variable specifying the rows in the table.

COL=

The name of the single variable specifying the columns in the table.

The following parameter is optional:

COUNT=

The name of the variable that contains the cell counts. This only needs to be specified if the input data set is already aggregated.

The version of the %RUN_FISHERS macro that you are using is displayed when you specify version (or any string) as the first argument. For example:

    %run_fishers(version, data=a, row=r, col=c)

LIMITATIONS:

The %RUN_FISHERS macro can only be used for a single table request. Specifying multiple rows and/or multiple columns (corresponding to a multiple table request such as (a b)*(c d) in the TABLE statement of PROC FREQ) is not supported.

No error checking is done. The macro assumes that the specified data set and variables exist and that their names are correctly specified.

Note that Fisher's exact test can become very computationally intensive as the sample size and the dimension of the table become large. This can result in extremely long run times and very large memory requirements. For details, see "Statistical Computations: Exact Statistics" in the Details section of the FREQ procedure documentation in the SAS/STAT User's Guide.

---

These sample files and code examples are provided by SAS Institute Inc. "as is" without warranty of any kind, either express or implied, including but not limited to the implied warranties of merchantability and fitness for a particular purpose. Recipients acknowledge and agree that SAS Institute shall not be liable for any damages whatsoever arising out of their use of this material. In addition, SAS Institute will provide no support for the materials contained herein.

---

---

These sample files and code examples are provided by SAS Institute Inc. "as is" without warranty of any kind, either express or implied, including but not limited to the implied warranties of merchantability and fitness for a particular purpose. Recipients acknowledge and agree that SAS Institute shall not be liable for any damages whatsoever arising out of their use of this material. In addition, SAS Institute will provide no support for the materials contained herein.

EXAMPLE:

This example checks to see if at least 80% of the cells in a 2x3 table have an expected value greater than 5 and then runs Fisher's exact test when it fails this test.

   data test;
     input r c ct;
     datalines;
   1 1 12
   1 2 5
   1 3 6
   2 1 13
   2 2 2
   2 3 4
   ;
   
   /* Define the RUN_FISHERS macro */
   %inc "<location of your file containing the RUN_FISHERS macro>";
   
   %run_fishers(data=test,
                row=r,
                col=c,
                count=ct)

RESULTS:

The %RUN_FISHERS macro produces the following results since only 50% of the cells have an expected value greater than 5.

The FREQ Procedure Frequency Percent Row Pct Col Pct Table of r by c r c Total 1 2 3 1 12 28.57 52.17 48.00 5 11.90 21.74 71.43 6 14.29 26.09 60.00 23 54.76     2 13 30.95 68.42 52.00 2 4.76 10.53 28.57 4 9.52 21.05 40.00 19 45.24     Total 25 59.52 7 16.67 10 23.81 42 100.00 Statistics for Table of r by c Statistic DF Value Prob Chi-Square 2 1.3571 0.5074 Likelihood Ratio Chi-Square 2 1.3895 0.4992 Mantel-Haenszel Chi-Square 1 0.6518 0.4195 Phi Coefficient   0.1798   Contingency Coefficient   0.1769   Cramer's V   0.1798   WARNING: 50% of the cells have expected counts less than 5. Chi-Square may not be a valid test. Fisher's Exact Test Table Probability (P) 0.0513 Pr <= P 0.5732 Sample Size = 42

Right-click on the link below and select Save to save the %RUN_FISHERS macro definition to a file. It is recommended that you name the file run_fishers.sas.

Download and save run_fishers.sas

This MACRO automatically performs Fisher's exact test whenever the chi-square test results in a WARNING message regarding the validity of the test in an RxC table because of a large number of cells with expected count less than 5.

Type:	Sample
Topic:	Analytics ==> Categorical Data Analysis Analytics ==> Exact Methods

Date Modified:	2009-03-25 15:30:15
Date Created:	2009-02-27 13:37:53

Operating System and Release Information

Product Family	Product	Host	SAS Release
			Starting	Ending
SAS System	Base SAS	z/OS
		OpenVMS VAX
		Microsoft® Windows® for 64-Bit Itanium-based Systems
		Microsoft Windows Server 2003 Datacenter 64-bit Edition
		Microsoft Windows Server 2003 Enterprise 64-bit Edition
		Microsoft Windows XP 64-bit Edition
		Microsoft® Windows® for x64
		OS/2
		Microsoft Windows 95/98
		Microsoft Windows 2000 Advanced Server
		Microsoft Windows 2000 Datacenter Server
		Microsoft Windows 2000 Server
		Microsoft Windows 2000 Professional
		Microsoft Windows NT Workstation
		Microsoft Windows Server 2003 Datacenter Edition
		Microsoft Windows Server 2003 Enterprise Edition
		Microsoft Windows Server 2003 Standard Edition
		Microsoft Windows XP Professional
		Windows Millennium Edition (Me)
		Windows Vista
		64-bit Enabled AIX
		64-bit Enabled HP-UX
		64-bit Enabled Solaris
		ABI+ for Intel Architecture
		AIX
		HP-UX
		HP-UX IPF
		IRIX
		Linux
		Linux for x64
		Linux on Itanium
		OpenVMS Alpha
		OpenVMS on HP Integrity
		Solaris
		Solaris for x64
		Tru64 UNIX