Observation on a child's behavior was conducted by two different clinicians. Their analysis were compared and finalised. This is called inter rater reliability.
A similar test, arranged in different format, is given to the same person one week apart and the results are compared. This is another concept of reliability called test-retest reliability.
Validity refers to the extent of which a test measures what it is supposed to measure. A car speedometer is supposed to measure the speed of the vehicle, not the air pressure of the tyre.
There are two types of validity - internal validity and external validity.
Internal validity refers to what extent the the test measures what it supposed to measure within the parameters of various variables.
For instance, in Zimbardo deindividuation concept,
Case A : a group of anonymous people, encountering an obnoxious man,
Case B: a group on non anonymous people, encountering pleasant woman,
In Case A, did they act aggresively due to being anonymous or due to the obnoxious man, or both?
In Case B, They didn't act aggresively, was it because they are non anonymous, or because there was a pleasant lady or both?
So internal validity is very concerned with what you are supposed to be measuring.
Just like a speedometer of a car is not supposed to measure air pressure even though it produces reliable and consistent figures.
External validity, on the other hand, concerns with how much extent the test in the laboratory setting can be generalised into a real life circumstances.
Delivering electric shocks in the laboratory setting is quite different from obeying the authority to carry out evil acts in the dark night.
No comments:
Post a Comment