# Types of errors in statistics.

##### Swamped with your writing assignments? Take the weight off your shoulder!

Type I and type 11 errors

Statistics has provided a venue for people to defend theories that were previously regarded as assumptions. However, even for statistics, the conclusion reached is not always 100 % correct. Why? Statistics cannot provide 100 percent confidence due to type I and type II errors. A hypothesis is a theory being tested. There are two types of hypotheses null and alternative hypotheses. If the theory is true, we reject the null hypothesis, but if it isn’t, we accept it.

Because of chance, there is always a possibility of reaching an incorrect conclusion in statistics, even if the correct procedure was used. Type One error occurs when a researcher rejects a null hypothesis when it is actually correct. Alpha which in most cases 0.5 represents the probability of making a type one error. The alpha value means that a theory being investigated would happen in a normal situation, but because of chance or a factor that rarely occurs the theory will fail or is incorrect. You correct a type one error by reducing the p-value; however, this would cause a type two error.

A type two error is when a researcher fails to reject the null hypothesis when it is actually false; a false negative. The researcher’s results indicate that there is no significant effect when it actually exists. Beta represents the probability of making a type two error. Beta is related to power, to avoid a type to error the test needs sufficient power. Sufficient power is when the sample size or effect size is large enough to run the test or decent a significant effect when it truly exists. When you reduce power, you increase the probability of committing a type II error, when you decrease power you increase the change of making a type I error.

Imagine a researcher investigating a drug and finding it is 95% effective. However, the drug is used on a patient with a rare condition and ends up not working because of a rare condition that the researchers did not consider in the investigation; this is a type one error. To correct the mistake, the researchers increase the rare condition cases (disregarding the condition is rare); and found no significant effect to prove the drug works, a type II error. The drug works, but to a 95% level, we can never rule out chance; therefore, we always have to accept a type 1 error.