W4 - test development Flashcards
Test conceptualisation
- Tests are developed to meet a particular need
- Your definition guides test development
- Need to identify all relevant content areas (dimensions/sub-domains) of the construct:
- A test typically samples all of these content areas;
- > i.e. samples the construct ‘universe’
- > without doing this, the test may lack content validity
What to determine during test conceptualisation
- purpose of the test (measure trait, or measure ability)
- audience of test (clinicians, educators, HR, researchers)
- type of sample (children, adults, language proficiency)
- method of administration
- > paper and pencil test vs. online administration
- > single versus group administration of test
- > response format of test (e.g. true/false, multiple choice, rating scale, etc.)
Rational driven approach
is guided by theoretical knowledge of the construct (e.g. test blueprint covers, domains a-b-c)
Empirical approach
to test development prioritizes ability to discriminate between two groups (e.g. clinical or healthy controls)
- > Item is good if clinical group scores high and healthy controls do not
- > MMPI-2 is an example of an empirical test that uses criterion validity
Types of data
Nominal data is categorical, and has no inherent “value”
Ordinal data is a ranking (1st, 2nd, 3rd , etc)
Interval data are measurements that provide information about order and magnitude of difference, i.e., there is meaningful distance between numbers
Ratio data has a meaningful true zero point
Examples of nominal data
- employment status
- gender
- ethnicity
Examples of ordinal data
- likert scale
- Olympic events
Examples of interval data
- temperature measures
- notes on a piano
- distance in calendar years
Examples of ratio data
- response time
- number of siblings
- savings
- height and weight
Item characteristic curve - ICC
A graphic representation of item difficulty and discrimination.
