Content area
Full Text
In the study by Finn and colleagues (2016), the researchers used a Fisher's exact test (full name Fischer's exact test of independence) to test the significance of the statistical comparisons. Fall rates before implementation of the continuous peripheral nerve block (CPNB) program for pain management following knee and hip arthroplasty were compared to falls after the program was implemented. This statistical test is useful for categorical data. In this column, the test and its importance will be explained.
Categorical Data
When data are continuous, such as a score on a scale or test, means are used to compare data from two or more groups. However, when data are categorical, means will not work. Categorical data, also called nominal data, are an either/or type of data: males or females, fell or did not fall, patients had a total knee arthroplasty or a total hip arthroplasty (TKA/THA) or they did not. This test examines if the null hypothesis that there is no difference between the results is true. In other words, the use of CPNB would have no influence on the number of falls that occur in the second time frame.
It is called an exact test because it identifies exactly the difference from the null or no difference hypothesis; other tests such as the chi-square are approximations (McDonald, 2014). Thus the Fisher's exact test often is used with small samples because it is more accurate than the chi-square test. Chi-square is used frequently with larger samples when results should...