Reliability Flashcards
(12 cards)
1
Q
What is reliability?
A
refers to how consistent a measurement is e.g. experiments, observations etc
2
Q
What are the two ways of assessing reliability?
A
- test- retest
- inter-observer reliability
3
Q
What is test-retest?
A
- method of assessing reliability of a questionnaire or psychological test by assessing the same person (or people) on two separate occasions
- if the test or question is reliable > the results obtained should be the same/similar each time they are administered
4
Q
What is the criteria for a test-retest?
A
- sufficient time between test and retest to ensure that p’s/respondents cannot recall their answers to Q
- but not so long that their attituded, opinions or abilities may have changed
- scores then correlated (e.g. spearman’s Rho) to make sure they are similar > if correlation is significant then the reliability of the measuring instrument =good
5
Q
What is inter-observer reliability?
A
extent to which there is agreement between two or more observers involved in an observation of behaviour
6
Q
How is inter-observer reliability established?
A
- pilot study of observation
agree on behaviour categories used - observers watch the same event but record their data independently
- data collected by observers should be correlated (e.g. spearman’s rho) to assess reliability
7
Q
How would you test reliability in a content analysis or interview?
A
- inter-rater reliability (content analysis)
- inter-interviewer reliability (interviews)
8
Q
How to measure reliability?
A
- using a correlational analysis
- two scores correlated
- correlation coefficient should exceed =.80 for reliability
9
Q
How to improve the reliability of questionnaires?
A
- pilot study to check if questions are misleading or complicated
- if some are complicated may replace open questions with close-ended ones > may be less ambiguous
10
Q
How to improve the reliability of interviews?
A
- same interviewer each time
- or interviewers must be properly trained so questions are not misleading or complicated
- pilot study to check if questions are misleading or ambiguous
- may just do a structured interview- fixed questions
11
Q
How to improve reliability of observations?
A
- observers must be well trained
- behavioural categories have been properly operationalised (measurable & self-evident) discussed & agreed
- way they collect & record data should be standardised
- conduct a pilot study to see if observers understand behavioural categories
12
Q
How to improve reliability of experiments?
A
procedures must be the same every time - standardised procedures
