RELIABILITY

→ how consistent a measuring device is

~INTER-OBSERVER RELIABILITY~

  • Tries to eliminate subjectivity and bias

  • Two or more independent observers watch the same event and analyse it using the same behavioural categories. If their data is similar (assessed through a correlation), then it is reliable

~TEST-RETEST RELIABILITY~

  • Used to assess the reliability of measuring tools

  • Involves giving the same participants the same test on two separate occasions. If their answers are the same, the test is reliable

~CORRELATION COEFFICIENTS~

  • In both inter-observer and test retest reliability, the two scores are created to form a correlation coefficient (-1 - +1)

~IMPROVING RELIABILITY OF QUESTIONNAIRES~

  • Questions deleted or reworded (if they’re too vague or confusing)

  • Some open questions replaced with closed questions (to make misinterpretation less likely)

~IMPROVING RELIABILITY OF INTERVIEWS~

  • Use the same researcher each time

  • Use structured interviews

~IMPROVING RELIABILITY OF OBSERVATIONS~

  • Behavioural categories must be operationalised, measurable and obvious

  • Train observers to use categories consistently

~IMPROVING RELIABILITY OF EXPERIMENTS~

  • Ensure the procedure is standardised

  • Ensure the procedure is clear and easy to understand