the criterion and predictor are measured at the same time. ex. if you were trying to measure college aptitude and gave a test to college students and compared it with their current gpa
how well a test predicts a criterion ex. sat-->college gpa
create two tests that measure the same construct and correlate them. do they correlate? if so, it has high convergent validity. ex. self report, and test.
if two dissimilar constructs have a low correlation ex. tests of self esteem vs manual dexterity
the variable that the test is trying to measure. ie. intelligence
questions to ask:
the proportion of applicants who are selected
the proportion of people expected to succeed if selected at random. the real value of a test comes from comparing the hit rate with the base rate. use old records of the percentage of people who usually succeed 60%
when you got admitted but you didnt do well
when you got rejected but you would have done well
gives the likelihood that a person selected on the basis of the test will actually succeed. There is a different table for each base rate. use old records of the percentage of people who usually succeed 60% then use validity (gre compared with actual success)..when u use the table u arrive at 66% which is based on gre scores. is the extra 6% worth it?
Item Response Theory
a graph of probability, where the X axis is innate ability and the Y axis is probability of passing an item
three steps in test development
test construction standardization test revision
based on an underlying belief or rationale
rely on criterion keying for their validity (factor analysis)
Tendency for subjects to respond to questions or test items in a specific way, regardless of the content.
social desirability (marlow crowne)
stanford-binet 4th edition
level1: g level2: crystallized, fluid, short term memory
stanford-binet 5th edition
Includes 14 separate subtests:
•Verbal tests: Information, digitspan, vocabulary, etc.
•Performance tests: picturecompletion, block design, symbol search, etc.
ages 2.5-12.5 based on well articulated theory of intelligence: rational scale
Designed to minimize cultural bias
Attempts to separate fluid andcrystallized intelligence
Consists mostly of nonverbal items
Divided into mental processingsection (fluid) and achievement section (crystallized)
3 applications for item response theory
item bias analysis-seeing if items are different for males vs females equating-equating two different tests (gre vs sat) tailored testing-giving different people different questions based on ability. (computer adaptive)
the extent to which an item discriminates those who recieve high scores vs those who recieve low scores
what does a steep upward slope indicate?
that people who did well one the test(or who have high verbal ability) are much more likely to respond correctly to the item
advantages of group testing
Group Intelligence Tests: Advantages and Disadvantages
More convenient and Economical
Examiner cannot: establish rapport, determine student’s level of anxiety, Student might not understand instructions, might be distracted by other students
The extent to which a selection tool predicts, or significantly correlates with, important elements of work behavior
weschler, weschler for kids, stanford-binet kaufman children
multidimensional aptitude test sat act gre asvab, gatb
multidimensional aptitude battery
just like the weschler but online version. a group test
Want to see the other 32 Flashcards in test 2 psy 370?JOIN TODAY FOR FREE!