Reliability

• Reliability

  Measure of consistency of a test/experiment and if a particular measurement can be repeated then it is considered reliable.
• Internal reliability

  Ppt gives a consistent pattern of answers, no matter how the researcher has phrased the question. Checking internal reliability with split-half method where in an interview/questionnaire, you split the questions in half to compare them to eachother.
• Test-retest

  Administering the same test to the same people on two/different occasions and measure the extent to which the rankings change over time. Often used with questionnaires/psychologcial tests but can apply yo interviews. Must be a sufficient time between test-retest to ensure the ppt does not recall their asnwers - their beliefs/attitudes/opinions cannot change. The results are correlated to see if they're the same/similar.
• Inter-observer reliability

  Recommandation that there should be two obserers to reduce bias, subjectivity and unreliability. A pilot study of the observation should be done to check that the observers are aplying behavioural categories in the same way. Observers may have to watch the same sequence of events but record their data independently. The test-retest method is used to correlate both observer's date to measure reliability.
• Improving reliability in Questionnaires

  Test-retest method comparing two sets of data should produce a correlation that exceeds +.80. Questionnaires that produce low test-retest reliability may require some items to be deselected/rewritten like replacing open qiestiones with closed ones.
• Improving reliability in Interviews

  Using the same interviewer each time or at least use interviewers that were all similarly trained. Avoid leading and ambiguous questions. Conduct structured interviews where the interviewer has less personal influence
• Improving reliability in Observations

  Behavioural categories should be properly operationalised and not overlap. This is vital as observers may make their own interpretations/judgements on the observation and may end up with differing/inconsistent records.