Reliability + Validity

Consistency of a psych test/measure

Actual measure (result) = hypothetical true score + measurement error

When measurement error is low

Internal
Inter-rater
Test-retest

Degree to which all items in a multiple item measure (eg. the questions in a questionnaire) measure the same thing (eg. all questions measuring positive/negative mood in a depression questionnaire)

Degree to which different assessors independantly agree on an observation/judgement

Assesses temporal consistency
Degree to which scores are consistent from one administration of a test to another (the same test)

Whether the test actually measures the underlying contruct [see psych as a science]

Test
Study

The measurement being used

To the research design

Content
Face
Criterion
Construct

How well measures represent the contruct (don't ask about history in maths exam)

Whether measure seems valid (eg. to pps - if pps think an IQ test is not accuratley measuring IQ they will have less confidence and may answer differently/their scores may change)

How well outcome of test relates to second related measure (A levels good predictors of degree classification)

How well the measure assesses the construct as a whole (maths questions might not fully assess IQ as a construct as it's only one aspect)

Internal
External
Ecological
Conceptual

Degree to which experimental study is methodologically sound

How well results of a study can be generalised to other situations/populations/time periods/environments

Whether the research can be applied to everyday activities

How well research hypothesis relates to underlying theory it was designed to test (does hypothesis/study logically follow from theory)