Validity Flashcards
Types of validity:
- internal
- external
- concurrent (type of internal validity)
Internal validity (what is it?)
The extent to which a test actually measures what it intends to measure
In an experiment - whether changes in the DV can be attributed to the manipulation of the IV rather than other sources of error (Extraneous variables)
- e.g. a well-controlled lab exp has high control over many EVs has higher internal validity compared to a field experiment
- in non-experimental methods - the extent to which the tool measures what it intends to measure
How to measure and assess Internal validity
ASSESS: face validity
- less rigorous - look at the experiment or research tool and check whether a measure appears, at face value, to test what it claims to
- an expert in the field could look at the test and ascertain whether it measures what it intends to measure
- OR an expert could rate it on a scale of 1-10 based on how well the test measures what it intends to measure
Concurrent validity (what is it?)
Whether a new test produces a similar measure of a variable to an existing well-established, validated test
How to assess Concurrent validity
- collect ppt data using the method from own exp/research tool
- use the same ppt+ gather data from them using an existing/established method in the field that is known to be valid
- Spearman’s Rho to compare ppt results from own test + existing/established test
- if correlation os 0.8+ - strong positive - new test validity
How to improve internal validity:
- having high levels of control
- standardised procedure
- do a pilot study to identify and correct extraneous variables before the study takes place
- covert observation - removes demand characteristics
- single blind trial - removes demand charachteristics
- double blind - helps reduce investigator effects
- training researchers to prevent investigator effects, e.g. standardised script or tone of voice
- randomly allocated conditions so ppts cannot decide conditions and gets rid of researcher bias
- repeated measures design should be counter balanced to limit order effects
- ideally use matched pairs design
- outline clear behaviour catagories
- questionnaires rather than interviews as they give ppts anonymity which reduces demand characteristics
External validity (what is it?)
The ability to generalise findings beyond the investigation itself. Whether the findings are representative of:
- other settings - ecological validity (real life)
- other groups of people - population validity (target population and other people)
- different periods of time - temporal validity
How to measure and assess external validity: (ecological, population, temporal)
How to assess: ECOLOGICAL validity:
- check concurrent validity of results by replicating research in a realistic setting (e.g. Godden and Baddeley classroom vs exam hall)
How to assess: POPULATION Validity:
- check concurrent validity of results by replicating with other groups of people (e.g. Van Ijzendoorn’s cross cultural meta analysis of the strange situation)
How to assess: TEMPORAL validity:
- check concurrent validity of results by replicating research in another time (e.g. Perrin and Spencer’s replication of Asch in 1980s.)
How to improve external validity? (ecological, population, temporal)
ECOLOGICAL: do research in a natural environment. conduct research in a realistic setting, use tasks which reflect real life activities (mundane realism)
POPULATION: conduct research with large, varied samples that reflect different subgroups of Target population
TEMPORAL validity: conduct research in different time periods - improve methods of the study if it doesnt replicate
