Types of validity and assessing and improving reliability Flashcards
What is validity?
Validity refers to weather a measuring instrument or study measures what it claims to measure.
What are the 4 types of validity?
Internal validity- This measures weather results obtained are solely affected by changes in the variable being manipulated (the independent variable) in a cause and effect relationship.
External validity- Measures whether the data can be generalsied to other situations outside of the research environment.
ecological validity- This is a type external validity that focuses whether the findings of the study can be applied to other setting such as everyday life. An example is that it is hard to generalise the findings of a laboratory study to real life due to the changes in behavior.
temporal validity- another form of external validity which refers to how much the findings can be applied across time ie the findings of asch’s study may not be the same today as it was conducted in a conformist era so his study lacks temporal validity.
What are the two ways to asses the validity of a study?
face validity- this looks at if a test appears to measure what it says it measure ie a questionnaire looking at health may not have face validity if not all its questions are about health.
Concurrent validity- This is where the performance of the test in question relates to a test that is already trusted in the same field. ie compare it to data taken from that test as if they measure approximately the same thing results should correlate.
Improving validity in experiments
A control group can be used to see if the independent variable influences the dependent variable.
Single blind or double blind procedures can be used. For single blind the participants don’t know the what condition they are in while in double blind it is both the participants and the researcher. This reduces demand characteristics and investigator effects.
Using standardised instructions which is when all participants get the same instructions in the same way , this mimeses investigator effects and the participants all have the same idea of what to do.
improving validity in questionnaires
A lie scale can be used to check the consistency of participants responses, an example is two contradicting statements might be asked and you have to rate how much you agree with them.
You can also assure the participant the questionnaires will be anonymous reducing social desirability bias.
improving validity in observations
By reducing the impact of the researchers you improve ecological validity.
Covert observation makes sure participants behave naturally.
Using behavioural categories will reduce the effects of researcher subjectivity.
Qualitative data is seen as more ecologically valid then quantitive data. However this may make the results more subjective reducing validity.
Taking data from a variety of sources can improve validity. you can take data from interviews, observations and written reports, this is process called triangulation
