It is a common fact knowing that in real life and practical case scenarios such a strict following of rules and assumptions while trying to calculate statistically becomes an anomaly, for there may be too many variables with unexplained behaviour that cannot be calculated exactly and on precise terms.
This may be in the case while measuring the life age expectancy of one or even when trying to predict long term weather conditions. Certain exogenous factors go towards proving these hypotheses wrong a lot of times. Despite using what one may feel are accurate measuring techniques and pre-defined and ordained scales, the possibility of getting wrong is still very omnipresent. This is due to the very nature of reality, where not everything can be quantified into precise terms and continuum can never be completely and wholly achieved.
There can be any variety of mistakes or errors that may occur before or during the calculation process. From sampling errors, where the sample size may be misrepresented, or too small, to recording errors, or question formulation errors, and even prejudice errors. These may all form the framework that may at any point discard the validity of the given subject and its results by leading hypothesis errors, where the right hypothesis may be discarded in the favour of the wrong hypothesis.
To ensure that such mistakes and errors do not generally take place, or are controlled to somewhat humane extent, different approaches and procedures have now come into adoption. One of the most popular is the Pearson coefficient which measures and reassesses the validity of different statistical manoeuvres to assure that certain maintenance of stability is carried throughout.