Week 2 Flashcards
When developing a measure, what are the key 4 things to consider?
- Reliability
- Validity
- Standardisation
- Bias
What does reliability refer to?
An extent to which a measure produces consistent results for individuals e.g if you measure height with a ruler you want to do the same for everyone every time
What are some of the test techniques used to quantify the reliability of measures?
- Test and retest - involves an individual taking a test once, then after a certain amount of time they are
tested again - Alternate forms - two different versions of an assessment are created - both of which are meant to measure the same variable
- Split half - Similar to alternate forms. They involve taking a questionnaire, separating it into two halves and administering them in the same time
- Inter rater - The measure is an observable characteristic
What does validity refer to?
Extent to which a measure measures what it is meant to measure
What are the three types of validity?
- Construct validity - is measure x related to past measures of x (new measures correlate well with existing measures)
- Content validity - Captures all aspects of x
- Criterion validity - predicts what x should theoretically predict
What does standardisation refer to?
Being able to compare individuals to an appropriate group
What is bias referred to?
One group is favoured in the testing method compared to other
What are the three approaches that aim to define intelligence?
- Psychometric approach
- Information processing approach
- Contemporary theories of practical and multiple intelligences
Define psychometric approach?
The psychometric approach tries to identify groups of items in a test that correlate highly with one
another in order to discover underlying skills or abilities. Participants perform multiple tasks, strong
performance on some tasks is likely to predict strong performance on others.
What is the primary tool of the psychometric approach?
Factor analysis - A statistical procedure for identifying common element, or factors, that underlie performance across a set of tasks