Reliability

• Reliability refers to how consistent findings from an investigation or measuring device are. If they produce similar results every time, it is said to be reliable
• Ways of assessing reliability:

  • Test-retest
  • Inter-observer reliability
• Test-retest:

  • questionnaires and psychological tests
  • ask the same ppts the same questions on separate occasions
  • correlate scores
  • enough time between test and retest so ppts forget answers, not so much time that their opinions and intelligence change
• Correlation coefficient must be +.80 or above for it to be reliable
• Improving reliability: Questionnaires

  • rewrite/ remove questions that are correlated below +.80
  • rewrite complex/ ambiguous questions
  • change open questions to closed questions
• Improving reliability: Interviews

  • same interviewer
  • train interviewers
  • structured over unstructured
• Improving reliability: Experiments

  • equal conditions
• Improving reliability: Observations

  • operationalise behavioural categories
  • no overlapping in categories
  • all behaviours covered