The misleading p-value
- P-values are common indicators of significance in statistical research. They represent the probability of observing a given null hypothesis.
- P-values are widely used to assess statistical significance. However, p-values have limitations worth considering. Two major ones are:
- P-values are strongly influenced by sample size. The larger the sample size, the smaller the p-value, irrespective of effect or difference.
- P-values do not show the magnitude or practical importance of an effect. Since the appropriate significance level is subjective to each researcher, p-values can encourage human bias and misinterpretation.
These limitations can easily lead to type I errors, false conclusions, and compromised research findings if unchecked.
As a researcher, it is essential to understand the limitations of p-values and include more tools of statistical inference to spice your research.
Click through the slides for a quick recap on the limitations of p-values.