reliability

• ways of assessing reliability

  inter observer

  test re test

  inter rater
• inter observer

  • together decide on categories
  • independently obseve same people using same categories (tally)
  • together compare tallies (see if results consistent)
  • >0.8 correlation (using spearmans rho)= good inter observer reliability
• test retest

  • do test
  • time interval eg. 1 week
  • same test same participants
  • compare both tests
  • correlate results
  • >/0.8 correlation = good reliability
• to improve reliability in observations

  operationalise behavioural categories so they are clear

  or more practice using them so they can respond quicker
• how to improve reliability in experiments

  standardising procedures so they are exactly the same each time
• how to improve reliability self report

  reduce ambiguity - rewriting questions to ensure everyone will interpret it the same way
• inter rater reliability (content analysis)

  • together decide on themes/categories
  • independently rate same material using same categories
  • together compare
  • correlate
  • >/ 0.8 correlation co-efficient = good reliability
• what is reliability?

  • how consistent the results are , it is reliable if the research be

  replicated to achieve consistent results