I have decided to talk about type I and type II errors mainly because I always get confused about which way round they go, and seeing as they are going in be in the exam I thought it would be a good method to help me learn type I and type II errors before the midterm.

So what are type I and type II errors? Type I and type II errors are mistakes made by the researcher when reporting the findings of a study. The errors can have adverse effects on society and might even be damaging.

Type II errors are errors made by the researcher who claims that they have found a difference when there was actually no effect. Type II error is when the researcher fails to reject a null hypothesis that is true. This is the worse out of the type of errors as it can have the worse effects on society. Type I errors do not have such adverse effect on society as these are mistakes where there is an effect by the researcher has failed to find it and has regarded it to not have an effect. In this case the researcher fails to reject a hypothesis that is false.  ( Nayman, Pearson)

An example of type I error in society could be damaging. For example, on non-psychological bases if a prisoner who was guilty of murder was found to be innocent then this would affect society. The murder has got away free and might be likely to re-offend again causing an innocent death again due to the wrong decision being made. This can be similar to a type I error in psychology as the jury have found a guilty man to be innocent, this might have occurred due to the juries not going over the evidence again to find the guilty person to actually be guilty. Type II error might occur due to this reason, the researcher may not check the data again or look at it carefully causing a studying to be released saying there is an effect when there is not.
Another example of type II error is when the researcher reports that there is a significance between the treatment of a particular drug and a placebo. If the researcher was to say that the drug was to have an effect and help the individual then again it could have adverse effect on society.  If the treatment drug had no impact on the individuals health it might cause depression as they might feel demotivated as a drug that was told that would help them has not in the slightest may cause negative thoughts an low self-esteem.

Type I error is when an effective is found but the researcher claims there is no effect, he has been able to find it. This may be due to the size of the effect or due to extreme outliers not being removed.  Again the researcher may have failed to look thorough the data finely before presenting the results. This may not have as much of an impact on society as the type II error. However it might have more of an impact on the individual, going back to the murder example the wrong person may have been found guilty. This may have negative effects on the individual as they have been found guilty for something they did not do. This may have occurred as the individual was at the wrong place at the wrong time and the evidence for the being may have caused the particular individual to look guilty. Another example of type I error is that there is a relationship between a treatment and placebo drug the finding might tell the reporter that there is no significance when there may have been a fault with the way it was explored. Again this would be due to the researcher not checking the data again. This would not have such an impact on society as it would not have negative effects on them as they have been told the treatment does not work. Therefore not causing any physical or mental harm to individuals receiving the drug.

To conclude type I and type II errors should be avoided when reporting results, data should be checked and ensured that no mistakes have been made. Type I errors can caused bad effects to individuals and therefore measure should be taken to ensure that data is not presented incorrectly.  I hope that I have got them around the right way.

 

http://www.experiment-resources.com/type-I-error.html

http://www.lifesci.sussex.ac.uk/home/Zoltan_Dienes/inference/Neyman%20Pearson.html

Advertisements