|
|||||
|
The reliability of a method of measurement, whether it is an experimental test, questionnaire or observation, refers to how consistently it measures. Thinking about reliability, there are two different types: Internal Reliability and External Reliability. Internal Reliability This type of reliability is concerned with how consistently a method measures. The important word in the previous sentence is "consistent", any test should produce consistent findings. For example, any test made up of half easy questions and half difficult questions would produce consistent results (everyone would score 50%) but this consistency is not the type of consistency we are looking for. This type of reliability is measured by 'split half reliability'. In this method you take your findings and the first half of your findings should correlate with the second half, e.g. the even numbered participants' findings should correlate with the odd numbered participants findings. External Reliability This type of reliability is concerned with how consistent a test is over time. Any test should give the same results if carried out on the same people on two separate occasions, each time under the same conditions. (If you did this you would have to be aware of "practice effects", the fact that someone has done a test once may mean that they improve on second time of testing.) The way to test this type of reliability is to use test re-test reliability. This involves testing a group of participants, re-testing them, but leaving sufficient time so that they forget about the test, and test them again. |
|||||
| Designed by Tony:Powered by IONOS |