reliability

Question 1

What is the strength of a corrlelation needed to be reliable ?

Answer 1

(+0.8) or more

Question 2

Define reliability

Answer 2

Refers to how consistent the findings from an investigation are.

Question 3

What are the two ways of testing or assesing reliability?

Answer 3

A method of measuring the reliability of a test overtime.

• If two sets of scores are similar then the test likely to be reliable not due to randomness
• Looking for +0.8 reliability

Question 4

what is meant by inter-rater reliability?

Answer 4

Refers to the consistency between observers involved.

• Independently complete the observation and come together and asses using a correlation to look for a strong + correlation (+0.8) has ^

Question 5

How is reliability quantified?

Answer 5

Reliability is quantified using the correlation coefficient 0-1 ( with closer to 1 meaning more reliable)

Question 6

How can a questionnaire improve reliability?

Answer 6

Measured with the test - retest method

• Q may need to be rewritten
• Avoid ambiguity / subjectivity
• Replace open questions (more room for misinterpretation) with closed questions (less ambiguous)

Question 7

How can an interview be improved to ^ reliability ?

Answer 7

• Use the same interviewer each time
• All observers should be trained / given a script - to ensure standardisation
• No leading questions/ambiguous
• More easily avoided in structured interviewers - behaviour controlled

Question 8

how can experiments be improved to ^ reliability?

Answer 8

Lab
* Standardised controlled conditions - to avoid variations
* Same instructions to all ppt
* Experimenter behave the same way → give them a script

Question 9

How can observations/ content analysis be improved to ^ reliability?

Answer 9

• Making sure behavioural categories are operationalised - measurable / less open to interpretation
• Should not have overlapping categories
• Train them on the system / coding process