Type-I and Type-II errors

When drawing conclusions, four scenarios are possible:

• Correct decision: the null hypothesis is incorrect, and the researcher rejects the null hypothesis.

• Correct decision: the null hypothesis is correct, and the researcher does not reject the null hypothesis.

• Type-I error: the null hypothesis is correct, but the researcher rejects the null hypothesis. The researcher falsely assumes that the independent variable had an effect. The chance of making a type-I error is called alfa level. It is common practice to use an alfa level of 5%. This implies that the null hypothesis is rejected even though there is a 5% chance that the found differences between the group means are caused by error variance. Thus, there is a 5% chance that the researcher wrongly rejects the null hypothesis. Sometimes, researchers use a stricter alfa level, for example an alfa of 1%. Then, they only have a 1% chance of wrongly rejecting the null hypothesis.

• Type-II error: the null hypothesis is wrong, but the researcher does not reject the null hypothesis. Thus, the researcher assumes that the independent variable did not cause an effect, while in fact it did cause an effect. The chance on a type-II error is called beta. Unreliably measuring the dependent variable raises the beta. Effects that do exists, are namely noticed less often with a larger beta. In addition, mistakes in collecting and coding the responses, highly heterogeneous samples and bad experimental control may cause an increased beta. To reduce the chance on a type-II error, researchers try to design studies with a high power.