Flashcards
Reliability and validity
Cards in this deck (14)
- 1Inter-rater reliabilityIn observations, at least two observers need to ensure that the data collected is reliable. Data from observers can be compared and statistical tests are conducted to assess the level of agreement. There is reliability if there is a strong positive correlation of 0.8 or above (80% conductive).
- 2Internal reliabilityThe measure used must be clearly outlined e.g self reporting rating scales (1-5) question- participants must be told what '1' means, what '5' means etc, So that there is no misinterpretation.
- 3Split-half methodAssesses internal reliability e.g. 1/2 items on a test/method correlated with the other 1/2. A strong positive correlation indicated internal reliability.
- 4External reliabilityConsistency over time when repeated.
- 5Test re-testAssesses external reliability. Gives to the same participants on 2 occasions and correlates the results from both. Strong positive correlation= good external reliability.
- 6Confounding variablesMay affect the dependent variable and therefore reliability. All variables need to be controlled.
- 7High reliabilityLab experiments (high levels of control/ replication/standardized procedures). Self reports = standardized & highly replicable. Observations (if inter-rater reliability is used and is over 0.8 or above).
- 8Ecological validityThis research is as close to real life situation/ behaviour as possible = high ecological validity e.g. participant observation + case study field experiments. If it is removed from real life, it is then low ecological validity as it would be controlled.
- 9Construct validityWhen a test or experiment measures the actual behaviour it sets out to measure.
- 10Criterion validityHow much one measure predicts the value of another one.
- 11Face validityOn the surface does the task or question appear to measure what you aimed to measure?
- 12Internal validityExperimental validity- to what extent are the results due to the manipulation of the IV? Extraneous variables need to be controlled as confounding variables may affect reliability and therefore validity.
- 13External validityTo what extent the findings can be generalised to other environments (ecological validity) and people (population validity)? Researchers will use random sampling and large sample sizes to increase the likelihood that their study will have high external validity.
- 14Concurrent validityComparing a new method or test with an already established one that claims to measure the same variable. A high positive correlation should be gained from the results of both tests.
Self reports may have low validity as responses may be influenced by social desirability and may be untrue. Therefore, results may be invalid.
Information
- Description
- Flashcards on Reliability and validity, created by Elisha Mottershead on 28/04/2017.
- Created by
- Elisha Mottershead
- Views
- 1
- Rating