• p value
• statistical significance
• postgraduate
• confidence intervals
• Cohen’s d

We are gradually moving to a world beyond ‘p<0.05’, where validity of the scientific conclusion of a research would be based on more than just the statistical significance itself. A large number of statisticians and scientists have called for an end to judging studies merely by statistical significance. The dichotomisation of p value is very much black and white; researchers should instead embrace the idea of the statistical grey. As postgraduate students, we tend to spend more time with statistical software packages, trying to get a statistically significant result, rather than focusing more on other important aspects such as study design, conduct of the study, systematic data collection and application of appropriate statistical models for data analysis and making inferences. Statistical inference is not, and has never been equivalent to scientific inference. Hence, we should report the exact p value as obtained from the data, without bucketing it as <0.05 or >0.05. After all, p value is a probability value and varies every time the study is repeated. We should be ready to embrace the uncertainty involved in scientific research instead of bringing in the false certainty of statistical significance. Therefore, whether the p value is less than an arbitrary threshold or not should not be the criterion while deciding which results to present or focus on.

In our postgraduate thesis work, the p value often becomes the ultimate measure of the significance of study outcomes. First, we should understand what the p value is and what it is not. Statistically, the p value refers to the ‘the probability of obtaining, by chance alone, a result at least as extreme as the one that was actually observed in a clinical experiment or epidemiological study, given that the null hypothesis (no difference between the specified population) considered is true’.¹ The p …