P-Value demystified

Abstract

Biomedical research relies on proving (or disproving) a research hypothesis, and P value becomes a cornerstone of “null hypothesis significance testing.” P value is the maximum probability of getting the observed outcome by chance. For a statistical test to achieve significance, the error by chance must be less than 5%. The pros are the P value that gives the strength of evidence against the null hypothesis. We can reject a null hypothesis depending on a small P value. However, the value of P is a function of sample size. When the sample size is large, the P value is destined to be small or “significant.” P value is condemned by one school of thought who claims that focusing more on P value undermines the generalizability and reproducibility of research. For such a situation, presently, the scientific world is inclined in knowing the effect size, confidence interval, and the descriptive statistics; thus, researchers need to highlight them along with the P value. In spite of all the criticism, it needs to be understood that P value carries paramount importance in “precise” understanding of the estimation of the difference calculated by “null hypothesis significance testing.” Choosing the correct test for assessing the significance of the difference is profoundly important. The choice can be arrived by asking oneself three questions, namely, the type of data, whether the data is paired or not, and on the number of study groups (two or more). It is worth mentioning that association between variables, agreement between assessments, time-trend cannot be arrived by calculating the P value alone but needs to highlight the correlation and regression coefficients, odds ratio, relative risk, etc.

Keywords

• confidence interval
• hypothesis testing
• non-parametric data
• null hypothesis
• null hypothesis significance testing
• parametric data
• p value