L3 - Reliability: Assessing & Improving Reliability

Question 1

Reliability

Answer 1

• based on consistency of measurement
• piece of research can be deemed reliable when the same research is carried out again in the future & produces the same results
• shows that the research & findings gained are consistent and the research can be viewed as good quality & trustworthy

Question 2

Internal Reliability

Answer 2

• concerned w/ whether the measuring instruments used in research can give the same results on different occasions
  e.g. ruler is a consistent objective measuring instrument but a rating scale may be more subjective and less consistent - therefore has less internal reliability

Question 3

Intra researcher reliability

Answer 3

• examines the consistency of the individual researchers behaviour during research
• intra means ‘within’ and intra research reliability is achieved if the researcher behaves consistently during research
• can be assessed by measuring the extent to which the researcher gains similar results when observing or measuring the same/similar situations more then once

Question 4

Test Re-test

Answer 4

• the same test is given to the same participants on two different occasions to see if the same results can be gained
• the scores from both tests are then correlated to see if they are similar
• if a positive correlation coefficient of +0.8/higher is found then the results are judged to be reliable

Question 5

3 methods of investigation

Answer 5

• observations
• self-reports (interviews/questionnaires)
• experiments

Question 6

Observations (assessing reliability)

Answer 6

test retest
- repeat observation again w/ same participants, compare results gained in both experiments (easier if observations are recorded)
- results should be similar - correlation coefficient +0.8
pilot study
- conduct a small trial, can ensure procedures/resources used in research can improve precision
- can incl. standardised instructions, debriefing & planning procedures properly, will minimise human error & variation
- all key terms need to be operationalised so researchers know what they are defining, measuring & observing
- operationalise behavioural categories & make sure all observers have been trained to look for appropriate behaviour to fit into each category

Question 7

Observations (improving reliability)

Answer 7

inter observer reliability:
- make sure observation isn’t biased
- use more then 1 observer & record behaviours separately, results should be compared and have correlation coefficient of +0.8
- if it’s low, can improve by ensuring the behavioural categories have been operationalised properly so that each observer understands the categories properly, observers may need further training
pilot study
- if results from pilot study aren’t clear, reliability can be improved by giving more training & practice to observers, they can become more familiar with the behavioural categories & respond more quickly when observing participants

Question 8

Self-reports (interviews/questionnaires) (Assessing reliability)

Answer 8

• when self-reporting, answers may differ on diff occasions depending on various factors, can assess reliability by:
  test retest
• give self-report to grp of participants, then repeat after some time (few weeks)
• compare the results of both self-reports, should be very similar with a correlation coefficient of +0.8 if self-report is reliable

Question 9

self-reports (improving reliability)

Answer 9

questions used in the interview
- make sure questions aren’t ambiguous and are clear so ppt understand them and would give same answers if q’s were repated
- need to remove/rewrite ambiguous questions
inter researcher reliability
- make sure self-report isn’t biased, if interviewing use +1 psychologist separately or together & record answers separately
- researchers need to act consistently and carry out the procedure and design in exactly same way to make research consistent
- if interview reliable, compare results and correlation coefficient +0.8

Question 10

experiments (assessing reliability)

Answer 10

• experiments are reliable if psychologists can conduct the experiment again & gain the same/similar results
  test retest
• conduct experiment & collect results, repeat few weeks later with same ppt in exact same way
• compare results - should be similar - correlation coefficient +0.8

Question 11

experiments (improving reliability)

Answer 11

Standardisation of instructions
- if an experiment is conducted twice, then repeat procedures twice
- procedures should be exactly the same for each ppt that participates, helps ensure results are reliable
- standardised instructions should be used & key concepts/variables should be operationalised