Common Misconceptions and Misunderstandings in Magic Cut-Off for Significance: P-Value

Abstract

- The p-value is the most commonly used statistical term to make a decision in medical statistics. It helps the researchers to decide whether the results (which are obtained from a hypothesis test) show statistical significance or not. Most researchers up today prove their decisions based in p-values. Although the p-value is used so commonly by the researchers, they are in high rate misunderstood, miss-or over-interpreted and also wrongly reported. The purpose of this study is to emphasize some of the misconceptions about p-value, and correct their misunderstanding and suggest to researchers a straight way to use p-value. Fisher, who is called as “father of statistics” were not actually the first one, who used the p-value, however he was the first to outline formally the logic behind its use. Fisher's defined for the p value as we use today: it is equal to the probability of a given experimental observation, under a null hypothesis. If this number were smaller than the acceptable threshold, researchers could "reject" the null hypothesis as unlikely to be true. The use of a threshold p value as a basis for rejection was called a "significance test." This is important to distinguish from the "hypothesis test," which will be discussed shortly. In our study we stated some of the misconceptions such as obtaining p-values from inappropriate statistical methods, p-values <0.05 shows clinical significance, using always two-sided p-values, a scientific conclusion is always based on “1” p-value, p-value is the probability of null hypothesis is true, format for table of p-values. As a result, some of the common misconceptions are highlighted about p-value. Being aware of these misconceptions, can increase of the quality of studies.