Content area
Full Text
Research conducted in prehospital, emergency public health, and disaster medicine is evolving. More comparative studies are published in these areas. Results are often reported in terms of means or medians to describe an effect for an intervention or event within a study population. Statistical tests applied within many of the studies are often reported in terms of a P value to associate findings with “statistical significance.”
“Statistical significance” is derived from hypothesis testing and is one research standard for determining the importance of an effect. Hypothesis testing generates P values. Simplistically, hypothesis testing is a process of comparing the probability for an effect of an intervention (study hypothesis) within a test population versus no effect (null hypothesis) within the same population. The P value is the probability that no effect of an intervention (null hypothesis) has occurred within a population. A large P value (higher probability for no effect) indicates no influence of an intervention. A small P value indicates that an intervention did have an effect. Traditionally, a P=.05 (5% probability of no effect) has been considered the cut off for determining statistical significance. Stated more precisely, by convention, a P=.05 or smaller indicates those data support an effect of an intervention upon a population and a P>.05 supports no effect of an intervention.
A statistically significant P value of .05 or less may or may not be of true clinical importance. In addition to data showing statistical significance of an effect on a population, the magnitude of an...