In the previous sections, we analyzed the data to identify potential root causes. Since there is a lot of effort and capital at stake as part of the improvement project, it makes a lot of sense to validate and confirm these root causes. We need to ensure that the results that we observe are not due to chance alone. Hypothesis testing is one such way to do this. A hypothesis test calculates the probability, p, that an observed difference between two samples being compared can be explained by random or chance variation rather than any real, non-random, or significant difference between the underlying populations from which the samples have been picked up. If this p value is small, typically less than or equal to 0.05, we conclude that the samples are drawn from different populations, and hence the change due to improvement is real and significant.

This principle can be used to: