Why is it important that statistical tests be utilized to analyze experimental results?

Answer 1

Statistical tests are needed to allow repeatability. This allows different products/systems/etc. to be tested using the exact same procedures and criteria; thereby allowing accurate comparisons.

Answer 2

With an experiment you have an idea, called the null hypothesis, which may in fact be something you do not believe is true, but whatever the case, with a given null hypothesis, and a properly designed experiment you should have a expected outcome.

The experiment's actual result usually varies due to a multitude of random effects including measurement errors.

This variation in outcome results will have a probability distribution associated with it. Extreme values from the expected result are associated with low probabilities, and are a good reason to reject the null hypothesis.

Statistical tests involve answering the central question about the probability of the observed outcome given the null hypothesis is true.

Answer 3

It's important because variation for different persons is hard to measure. Therefore, you have to use statistics to be able to express the differences as to how to measure experiments which can be very susceptible to high variation.