P-values, short for "probability values," are statistical measures used to assess how strongly the evidence contradicts a null hypothesis.
Distribution of P-Values
P-values are obtained from random samples, so they follow a certain pattern of distribution.
When the null hypothesis is true, all p-values between 0 and 1 are equally probable.
However, if there's an actual difference (alternative hypothesis), p-values tend to be closer to 0 rather than 1.
Interpreting P-Values
P-values are often misinterpreted. When we see a "significant" result, it means that our sample size in the model is big enough to give us confidence that the differences we've observed are not just happening by chance.
P-values are interpreted consistently across different tests.
Small P-Values:
A P-value less than a predefined significance level (often set at 0.05) indicates strong evidence against the null hypothesis. This means that the observed results are unlikely due to random chance alone, leading to rejection of the null hypothesis.
Large P-Values:
A P-value greater than the significance level suggests weak evidence against the null hypothesis. In such cases, there isn't enough evidence to reject the null hypothesis.
Here's a breakdown of interpretations based on P-value ranges:
-
P-value over 0.1: No convincing evidence against the null hypothesis.
-
P-value between 0.05 and 0.1: Very weak evidence against the null hypothesis.
-
P-value between 0.01 and 0.05: Moderately strong evidence against the null hypothesis.
-
P-value under 0.01: Strong evidence against the null hypothesis.
Examples
P-value = 0.0023
Interpretation: With such a small p-value, there's only a 0.0023 probability of obtaining these results if the null hypothesis were true. This suggests strong evidence against the null hypothesis.
P-value = 0.4
Interpretation: A p-value of 0.4 indicates a 40% chance of observing such results if the null hypothesis were true. Thus, there's no substantial evidence against the null hypothesis.