RM: Validity and Reliability Flashcards
What is reliability?
-if results are consistent (when same methods are used)
-reliable methodologies will produce similar results each time
What is internal reliability?
the consistency between measures in a study
What is external reliability?
If the measure is reliable when repeated
What is test-retest reliability?
-measures test consistency (over time)
-study is repeated using the same procedures and the correlation between the two results are tested
-0.8 or above means reliable
What is inter-rater reliability?
-the degree of agreement among raters/observers
-2+ observers record the same observation using the same behavioral categories
-then test the correlation of their results
What is validity?
-the extent to which something is measuring what it claims to
-valid methods measure the construct they are supposed to
What is internal validity?
-whether it is the IV that affects the DV
-if the DV is changed by other factors/uncontrolled EVs, there is low internal validity
What is external validity?
-questions if a study’s findings can be generalized to the outside world
What is ecological validity?
-setting
-the extent that the findings of a study can be generalized to other environments/real world
What is population validity?
-people
-the extent the sample used can be generalised to the whole population
What is temporal validity?
-time
-the extent the results of a study can be generalised to other time periods
-are past experiments still valid?
What is construct validity?
-if something is measuring what you think it’s measuring
What is concurrent validity?
-the extent that data from a new test is similar to the data from the old test of the same variable
What is predictive validity?
-the extent results can predict what will happen in the future
What is face validity?
-the extent the test appear to measure what it claims to
How to improve internal validity?
-random allocation
-standard procedures
-counterbalancing
-single & double blind trials
-peer review
How to improve external validity?
-replicating findings in multiple sample groups
How to use observations to improve reliability?
-improve the training given to observers (improves observation to the correct behavioral category)
-pilot studies can be used to identify weaknesses
How to use interviews to improve reliability?
-use structured rather than unstructured
-scripts with interviews
How to use questionnaires to improve reliability?
-closed questions (reduce range of responses)
-use established rather than new questionnaires
How to use experiments to improve reliability?
-use standardized procedures
-use established rather than new ones
What is a correlation coefficient?
-measure of how closely related variables are
-if there is strong correlation, the variables are closely related
