Reliablity and validity
- Created by: Hannah Jeffery
- Created on: 16-04-15 13:13
View mindmap
- Reliability and validity
- A method is reliable when different researchers can measure or observe something and get similar results
- External reliability
- this is the ability to produce the same results every time the test is carried out
- this can be tested using the test-retest method
- where the test is carried out several times using the same or similar participants
- may vary due to individual differences, extraneous variables etc
- Test of correlation between set of two scores
- high correlation coefficient indicates the test is reliable, low correlation coefficant indicated the test is unreliable
- this can be tested using the test-retest method
- this is the ability to produce the same results every time the test is carried out
- Internal validity
- consistency of a measure within a test
- ensure that your measure is testing accurately
- Checked by the split- half method
- comparing half of the test with the other to check whether the scores of a variable are consistent
- External reliability
- Improving reliability
- where possible take more than one measurement from each participant
- reduces the impact of any anomalous score
- pilot studies ensure that a measurement works properly and that participants can use any apparatus successfully
- to ensure that all researchers collect and record data in a standardised way
- for interviews and observations this may require a period of training
- directly improve inter-test reliability
- data should be checked thoroughly after being recorded
- to ensure data isn't written down and entered in the computer wrong
- where possible take more than one measurement from each participant
- Internal validity
- The ability of the study to test the hypothesis that is was designed to test
- Is the Dv being affected by confounding variables instead of just the iv
- the greater control we have over these variables the better the internal validity
- Is the Dv being affected by confounding variables instead of just the iv
- maybe affected by demand characteristics
- cues that lead to the tendency for participants to behave in the way they think is required of them
- eg experimenter bias
- cues that lead to the tendency for participants to behave in the way they think is required of them
- improving
- single blind technique
- participants don't know what condition they are in
- therefore cant change their responses to please the experimenter
- double blind technique
- neither experimenter or participant know which condition they are in
- single blind technique
- The ability of the study to test the hypothesis that is was designed to test
- assessing validity
- face validity
- a judgement of whether the test is valid
- weakest form of validity testing
- a judgement of whether the test is valid
- criterion validity
- whether the test produces similar results to other tests that measure the same thing
- objective
- two types
- concurrent
- shows similar findings to another existing measure
- predictive
- how well a test predicts future performance
- concurrent
- face validity
- external validity
- how well the results beyond the study can be generalised
- population validity
- the degree to which results can be generalised to the rest of the population
- ecological validity
- the degree to which findings can be generalised to other situations, places and conditions
- mundane realism
- degree to which the setting or procedure reflects real life
- experimental realism
- the degree to which the results reflect realistic behaviour
- internal v external validity
- the more confounding variables that are controlled the more artificial the study becomes
- researcher must consider which is more important for his study
- A method is reliable when different researchers can measure or observe something and get similar results
Comments
No comments have yet been made