When testing a hypothesis there are two possibilities: an effect and no effect. Accuracy is the most important element in psychology, therefore we only accept a result when there is 95% confidence in the effect and only 5% chance results could occur if there was no effect. However even if we’re 95% confident, there is still a chance we can get it wrong! Type One Error and Type Two Errors are the two mistakes we can make:
- A Type One Error is when we believe that there is an effect but in reality there isn’t one, this is known as the α level. It is when we wrongly accept the experimental hypothesis and reject the null hypothesis, when it is actually true. This is the worst type of error because false conclusions could do lots of damage e.g. incorrect diagnoses.
- A Type Two Error is the opposite; it is when we believe there is no effect when in reality there is. This is most common with small test statistics. The maximum acceptable probability of a type two error is 20%, this is called the β level. In summary, a type two error occurs when we wrongly reject the experimental hypothesis, and accept the null hypothesis when it is false. This is the best type of error to make because no damage is done, the only disadvantage is that nothing good comes of it either e.g. new findings.
An illustration of type one and type two errors in the real world is Rosenhan’s on being sane in insane places study (1973). The study aimed to test whether psychiatrists could reliably diagnose whether a person was insane or not. There was two parts to this study:
- In the first part of the study, pseudopatients arranged an appointment at the hospital and pretended to be insane. These patients were admitted to the psychiatric ward and labelled as insane, this is an example of a type one error because they were being diagnosed incorrectly.
- In the second part of the study, Nurses were told that pseudopatients would be trying to enter the hospital however this was a lie. Because nurses believed pseudopatients would enter the hospital, they failed to diagnose many patients that were really insane. A type two error was made because they failed to diagnose individuals.
From both parts of Rosenhan’s study, it is clear that both type one and type two errors can have a big impact in the real world. But is it more damaging to wrongly diagnose an individual or to fail to diagnose an individual? From the study it is clear that type one errors are the most damaging; diagnosing an individual wrongly can cause all sorts of problems such as hospitalisation, administration of medication and labelling, all of which have a profound negative impact on an individual’s life. Failing to diagnose an individual can have negative impacts for the individual and society, but this impact is not as negative as wrongly diagnosing a person is, as the undiagnosed individual can make another appointment for further diagnosis.
In conclusion, when considering Type One and Type Two Errors, I believe that type one errors are the most damaging in both research and real life because they cause harm rather than just failing to do good.
Andy Field Textbook – Discovering Stats Using SPSS – 3rd Edition
Rosenhan, D.L. (1973) On being sane in insane places. Science, 179. 250-58