Forskellig betydning af p-vaerdien i eksplorerende og konfirmerende hypotesetestning
Sammanfattning
The outcome of clinical studies is often reduced to the statistical significance of results by indicating a p-value below the 5% significance level. Hypothesis testing and, through that, the p-value is commonly used, but their meaning is frequently misinterpreted in clinical research. The concept of hypothesis testing is explained and some pitfalls including those of multiple testing are given. The conceptual difference between exploratory and confirmatory hypothesis testing is discussed, and a better use of p-values, which includes presenting p-values with two or three decimals, is suggested.