Reliability Flashcards
the key word when thinking about reliability is
consistency
external reliability refers to
the extent to which a measure is consistent over time
what is the test-retest method
giving the same test (or interview/questionnaire) to the same person on a second occasion and checking if the results are the same
improving external reliability - standardisation
use standardised instructions and procedures so all participants have the same experience
improving external reliability - control
all the possible extraneous variables so different extraneous variables don’t mean a lack of consistency
improving external reliability - measurement
use an objective scientific standardised measuring tool so there is consistency in measurement
how can researcher bias be linked to reliability
if there is researcher bias then the results wouldn’t be consistent between researchers over time
how could we control researcher bias
independent researchers, double blind
observations should be
consistent
the type of reliability relevant to observations is
inter-rater reliability which means each observer is consistent with each other
to check the inter-rater reliability between researchers/observers their observations should be
correlated
what is the general rule of reliability
there should be more than 80% agreement on the observed behaviours for the data to have inter-rater reliability
a self-report should be … with itself and over several different …
consistent, occasions
the types of reliability relevant to self-reports are
internal reliability and external reliability
internal reliability
a measure of the extent to which something is consistent within itself. e.g. all the questions on the questionnaire should be measuring the same thing
external reliability
a measure of consistency over time. e.g. if the same interview/questionnaire was conducted one day then a week later, the outcome should be the same
method for assessing (checking) internal reliability
split-half method: the two halves of the questionnaire can be compared. this can be done by randomly selecting half the test items and placing them on form a and placing the other half on form b. therefore, you end up with two forms of the same test
method for assessing (checking) external reliability
test-retest method: involves giving the same questionnaire/interview to the same person (by the same person) on two separate occasions (with a sufficient gap for the chance to forget). if both yield the same results then it was reliable
dealing with (improving) low internal reliability
remove questions that don’t appear to be getting consistent results
dealing with (improving) low external reliability
1) rewrite questions so they are not ambiguous or leading
2) train interviewers so they all deliver the questions in the same way and not encouraging the respondents answer in a certain way
