reliability & validity

Cards (13)

  • reliability- consistency of results
  • inter observer reliability:
    1. researchers have the same training and look for the same behaviours
    2. two psychologists independently observe
    3. they check for agreement
    4. a +0.8 correlation shows reliability
  • improving inter observer reliability:
    -more training
    -operationalisation (specific behaviours)
    -video the behaviour, and let them view the same video
  • test-retest reliability:
    1. the same person does the same test on two different occasions
    2. if they find similar results, it is reliable
    3. this could be done by a strong positive correlation of +0.8
  • improving test-retest reliability:
    -standardisation (instructions, materials, procedure)
    -identify and eliminate EV’s ( could use a lab experiment)
  • internal validity- a test is measuring what it is supposed to be measuring
  • internal validity- it is measuring what it is supposed to be and the findings are due to only the IV. so we can establish a cause and effect relationship.
  • improving internal validity:
    -minimise EV’s
    -Lab studies
  • ecological validity- generalised to the real world
    improving ecological validity:
    -more natural environment
    -use tasks/materials that represent every day life
  • population validity- can be generalised to the target population
    improving population validity:
    -large sample sizes
    -random/stratified sampling
    -conduct research cross culturally across different ages and genders
  • temporal validity- can be generalised to the modern day
    improving temporal validity- do the test again to see if results are still accurate
  • concurrent validity- close agreement between data produced by the new test and the established test. the two tests exceeds +0.8.
  • face validity- it has on face value and looks like it is measuring what it is supposed to.
    you can hand research to another psychologist to see if it has good face validity