Reliability/Validity

Cards (17)

  • What is reliability in psychology?
    A measure of consistency
  • How is inter-observer reliability assessed?

    By correlating the observations or scores of two or more observers
  • What correlation coefficient indicates reliability in inter-observer reliability?
    +0.8 or higher
  • What does test-retest reliability involve?
    The same people repeating a measure on different occasions
  • What correlation coefficient indicates reliability in test-retest reliability?
    +0.8 or above
  • What is validity in psychology?
    The extent to which something is true or legitimate
  • What is internal validity?

    Whether a researcher measures what he/she intended to
  • What is external validity?

    Whether results can be generalized outside of the sample or situation
  • What is face validity?

    Does the test appear to measure what it says it measures?
  • What is concurrent validity?
    The performance of the test is compared to a recognized test in the same field
  • How can reliability be improved in questionnaires?

    • Change open questions to closed questions to minimize ambiguity
  • How can reliability be improved in interviews?
    • Ensure the same interviewer conducts all interviews
    • Provide training to limit potential bias
  • How can reliability be improved in experiments?
    • Increase control over extraneous variables
  • How can reliability be improved in observations?

    • Operationalize behavioral categories
    • Ensure there is no overlap between categories
  • How can validity be improved in experiments?
    • Use control groups to see the impact of an independent variable
    • Implement standardized instructions and procedures
    • Use single-blind and double-blind procedures
  • How can validity be improved in questionnaires?
    • Include lie scales to check for consistency
    • Use two questions that ask the same thing in opposite ways
  • How can validity be improved in observations?
    • Use covert observations to reduce bias
    • Use operationalized behavioral categories that do not overlap