Type II Error

When Does Type II Error Occur?

When researchers make decisions on statistical testing results based on a p value, there is a chance that researchers might make errors such as Type II error by accepting a false null hypothesis. That is, the truth in a population is that a research hypothesis (alternative hypothesis), which predicts that there is a relationship between an independent and dependent variable, is true. However, data from a sample indicate that a p value is greater than the conventionally accepted level, so researchers conclude that the two variables in the research hypothesis are not related. By erroneously accepting an inaccurate null hypothesis and simultaneously failing to support a true research hypothesis, researchers make a Type II error. The probability of committing Type II error is called beta (β).

As an example, consider a situation in which a researcher hypothesizes that the amount of time spent on Facebook is related to friendship satisfaction. Consequently, a null hypothesis is that there is no relationship between the two variables. As hypothesized, the truth in a population is that there is a relationship between time spent on Facebook and friendship satisfaction, so the null hypothesis should be rejected. However, by chance, the level of significance from the selected data might not meet the conventionally accepted level of a p value. [Page 1804]Based on this finding, the researcher would conclude that time spent on Facebook has nothing to do with friendship satisfaction when they are in fact related. In this case, this researcher would end up committing Type II error by failing to reject a false null hypothesis.

Theoretically, the probability of committing Type II error can be reduced by statistical power. Statistical power indicates the likelihood of rejecting a null hypothesis when it is false and should be rejected. Thus, if statistical power is strong, the probability of reducing Type II error becomes high. Power can be assessed as: 1- beta (β), and it can be improved by increasing a sample size. A larger sample size leads to a stronger power. Ultimately, the likelihood of committing an error can be reduced.

Relationship Between Type I and Type II Errors

Type II error is interrelated to Type I error (another type of error in statistical testing). The more researchers try to reduce the likelihood of committing Type I error, the greater the probability of making Type II error becomes, and vice versa. More specifically, when trying to reduce Type I error, researchers would decide to lower a p value. Then, it naturally becomes more difficult to reject a null hypothesis, and consequently this would lead to a lower chance of committing Type I error. However, a smaller p value would make it harder to reject a null hypothesis even when the null hypothesis is false and should be rejected: Type II error.

...