Y13 RESEARCH METHODS - reliability

Question 1

what is reliability?

Answer 1

A consistent measurement

Question 2

what is internal reliability?

Answer 2

The extent to which a measure is consistent within itself.

Question 3

What is external reliability?

Answer 3

produces the same results each time tests carried out.

Question 4

What are some ways of assessing reliability?

Answer 4

Test-retest
Inter-observer reliability

Question 5

Explain the test retest method?

Answer 5

Repeat the test again, if test results are similar data should be reliable.
There should be sufficient time between test and retest so participants can’t recall their answers.
BUT not a very long time that their attitudes, opinions, abilities may have changed.

Question 6

Explain inter observer ( inter rater ) reliability?

Answer 6

In an observation = two or more research is compare the data by conducting a pilot study to check that observers applying behavioural categories the same way.
Observers should watch the same event or sequence of events, but record data independently.

Question 7

Measuring reliability

Answer 7

In test retest + inter observer reliability = two sets of scores are correlated.
In order to be deemed reliable the correlation coefficient should exceed + 0.8 for reliability.

Question 8

Explain how to assess internal reliability?

Answer 8

SPLIT HALF METHOD = splitting a test into two and having seen participants doing both halves of the test.

Compare results for the 2 halves.
If 2 halves of the test provides similar results = TEST HAS INTERNAL RELIABILITY

Question 9

How to improve reliability in questionnaires?

Answer 9

repetition of similar questions
use clear questions

questionnaire that produces low test-retest reliability = questions need to be re-written/removed
opened qs replaced with closed qs

Question 10

How to improve reliability in interviews?

Answer 10

Use the same interviewer + same questions each time.

Interviewers must be trained to avoid leading questions = if same interviewer cannot be used

Leading questions can be avoided in structured interviews.

Unstructured interviews are free-flowing + less likely to be reliable.

Question 11

How to improve reliability in observations?

Answer 11

• make sure that behavioural categories are operationalised and measurable.
• categories should not overlap
  -all possible behaviours should be covered on the checklist.
• low reliability = observers may need further training in using behavioural categories.

Question 12

How to improve reliability in experiments?

Answer 12

• standardised procedures = must be same every time.