Reliability - Methods

notes on reliability for methods

a2 psychology with aqa

HideShow resource information
  • Created by: Cara
  • Created on: 29-06-12 13:45
Preview of Reliability - Methods

First 310 words of the document:

Types of reliability:
External reliability
o Consistency between measures
o E.g. a ruler would always measure the same in different situations
o In psychology, an observational study would need to give the same results in
different contexts, for example, observing male/female children's play interactions
at school and at home
Internal reliability
o Consistency within measures
o E.g. using a stop watch to time different performances in a range of different tests
o In psychology, using a questionnaire on racial prejudice and making sure that each
item measures the same construct at the same level
There are a number of different ways to check for reliability
1. Inter-rater reliability
This looks at how closely different individuals score something
Frequently used in observational studies
Results are compared and there should be a strong positive correlation ­ especially if results
are objective such as a multiple choice test
If rating is subjective ­ such as an essay ­ there will be less inter-rater reliability
2. Test-retest reliability
This checks external reliability, i.e. does the test produce similar results on at least two different
occasions. To do this you would need to use the same people on each occasion, as well as there
being a gap between administering the two tests
There should be a correlation of over 0.7 between the scores
However, scores may be different or inaccurate due to participant variables such as
Order effects
Extraneous variables such as personal experience or situation
3. Split-half reliability
This measures the internal reliability, i.e. the extent to which each part of the test gives the same
result as other parts of the test.
Useful to check the reliability of a questionnaire

Other pages in this set

Page 2

Preview of page 2

Here's a taster:

For example:
In a questionnaire with 50 questions, either
Compare the responses of items 1-25 with those from items 26-50
Or, compare responses from odd numbered items with those from even numbered items
Correlations achieved would be expected to exceed 0.9
4.…read more


No comments have yet been made

Similar Psychology resources:

See all Psychology resources »See all resources »