Lecture 2- Psychometric Test Theory Flashcards
How might the intellectual future be predicted?
1) achievements of parents
2) progress after training has been given
3) identify prerequisites necessary for future success
4) general cognitive abilities
What should all intelligence test have?
- high reliability
- low standard error of measurement
- validity
What is reliability?
when a test measures a construct accurately so therefore the results are consistent
Describe test-retest Reliability
1) The same ppts are tested twice on 2 separate occasions
2) results from these tests are correlated
3) 0.7-0.8 is a good correlation range
4) accurate scores will be consistent overtime
Describe Inter-Scorer Reliability
1) used for tasks that have no clear marking criteria
2) same responses scored by 2 independent judges
3) look for a high correlation between paired rating score
4) accurate scores will be judged consistently
Describe Parallel-Form Reliability
- Test the same people twice with two different test versions
- Look for a high correlation between paired test scores * Accurate scores will be consistent amongst
different versions of the same test
What is Standard error of Measurement?
indicates how likely it is that test scores are good estimates
What does it mean if there is a high standard error of measure?
individual scores are not close to true scores, so scores discriminate poorly between candidate
What does it mean when there is a low standard error of measurement?
individuals scores more likely to be close with true score, so scores discriminate well between candidates
Define validity
test that measures what it intends to measure
Name the types of validity?
- Predictive
- Differential
- Concurrent
- Construct
What is predictive validity?
- Intelligence tests should predict success
for tasks that require intelligence - Test score and task performance should be correlated
What is concurrent validity?
- Intelligence tests should be highly correlated with
established benchmark measures of intelligence
What is construct validity?
- An intelligence test is valid if it demonstrates multiple
converging validities, for example, as described above
How do intelligence test differ?
1) types of tasks
2) The number of different tasks
3) The extent to which they attempt to be culture-neutral
4) Administered individually or as a group
What is an individual Intelligence test ?
- Taken face-to-face, with a trained administrator assessing one single individual
- best for diagnosis, educational setting
What are the advantages of an Individual intelligence test?
- Use a wide variety of tasks [reduces SEM]
- Motivation can be maintained (e.g., targeted difficulty) - Performance can be interpreted (e.g., near misses)
What are the disadvantages if an individual intelligence test?
- Administrator must have appropriate training
- Time-consuming to administer
What is a group intelligence test?
- Several people take tests simultaneously under the exam
conditions supervised by a (less-trained) individual - Best for screening, often used in occupational settings
What are the advantages for group intelligence testing?
- Relatively fast and efficient to administer
- Advanced training not necessary for an administrator
- Objective scoring, usually multiple choice
What test items could be found in an intelligence test?
1) Analogy (non-verbal/ verbal)
2) Sequence (non-verbal/ verbal, Gen knowledge, numerical)
3) Odd one out
4) Spatial manipulation
5) Matrix Items
Give examples of WISC items
- Picture concepts
- Matrix reasoning
- Bock design
- picture completion
What do the items focus on during group intelligence testing?
- focused more on logical reasoning
1) identify a rule or relationship from item elements, and apply the rule to generate the correct answer - focused more on the relationship between shapes and patterns
- high score finding 1 single correct answer for each item: limited creativity
