reliability

Question 1

inter-observer reliability:

Answer 1

several researchers are asked to analyse the same data. All have the same behaviour categories and when observing, do not interact. Results are correlated using Spearman’s Rho test, need a correlation of at least 0.8.

Question 2

test-retest reliability:

Answer 2

repeating the same test on the same participants to check for consistency. Usually a small interval in between, so the participants don’t remember their answers. Scores compared using Spearman’s Rho test (need 0.8 correlation)

Question 3

inter-interviewer reliability:

Answer 3

different interviewers interview the same participant with the same questions/test and the answers are compared using correlation coefficient of Spearman’s Rho test (need at least 0.8)

Question 4

improving reliability: behavioural categories

Answer 4

make sure they are clearly defined and operationalised, so disagreement about their meaning is less likely. As well as ensuring the observers are trained and practices in using behavioural categories.

Question 5

improving reliability: reducing ambiguity

Answer 5

• make sure the questions/task is very clear and not ambiguous.
• closed questions
• using the same interviewer each time.
• no leading questions

Question 6

improving reliability: standardisation

Answer 6

• keep everything standardised so that the procedures can be repeated.
• control confounding variables