Reliability

Question 1

Reliability

Answer 1

The extent to which a measurement of behaviour is consistent

Question 2

Internal reliability

Answer 2

-consistency within a study or measure (tool to measure behaviour / procedure and /or researcher)

Question 3

Issues of internal reliability

Answer 3

More than one researcher - is there consistency in how they are interpreting behaviour or carrying out the study?
More than one measure - if there is more than one test (such as used in a repeated measure design then is there consistency of challenge in each test?
Within a measure - If a measure has more than one question (eg. IQ test) or item (e.g. different words in a memory list or obstacles in a driving test) then are these questions/items consistently measuring the same thing?

Question 4

How to overcome internal reliability issues

Answer 4

How to overcome issues
- Clearly operationalise variables so there is less room for different interpretations.
Standardise procedures to ensure consistency
- Detailed training of researchers to ensure consistency
- Standardise the measures, i.e. make sure they are the same in difficulty etc.
- Standardise each item or question within the study, e.g. each word in a memory list has the same number of letters

Question 5

How to assess if internal reliability is reliable

Answer 5

How to assess/establish if reliable
Inter-rater reliability - involves ensuring that two or more psychologists produce consistent results within the research by using a standardised procedure, agreed coding system, or correlation of their data. If they use consistent measures and/or there is 80% agreement or above then the research has internal reliability.
Split-half - Split half method involves getting participants to complete the test to measure behaviour then once done splitting a participant’s test answers in half and seeing whether they got the same or similar scores on the two halves of the test. If so, internal reliability is high; if not, it is low and individual questions would need to be redesigned

Question 7

External reliability

Answer 7

Consistency outside study /across time (different occasions )

Question 8

Issues of external reliability

Answer 8

Behaviour is only measured once - the test may not measure behaviour consistently at other times, e.g. a personality test may only indicate Ppts personality traits at the time the study was carried out

Question 9

How to overcome external reliability issues

Answer 9

Ensure tests etc. are designed in relation to measuring behaviours, attitudes etc. beyond one setting or time

Question 10

How to assess wether external reliability is reliable

Answer 10

Test-retest - involves testing and retesting the same participants over time, with the same test/procedures, and comparing the results from one occasion to the next. If the scores are the same over time the test has external reliability.