exam 3 Flashcards
what are the 4 sources of measurement error?
• test taker
• test admin
• test itself
• test setting
reliability or validity - 1 test?
reliability
reliability or validity - 2 tests?
validity
does the test look like its measuring what its supposed to? - what is this measuring?
face validity
your score (X) on any test is never a perfect score, it’s always a combination of ____ (T) and ____ (E)
what you would have gotten if it was a perfect score (T), some kind of error (E)
we reduce error by increasing ____
consistency
if the test measures what it was designed to measure, it has ____
hint: reliability or validity
validity
test that are similar to what you’re measuring show ____ validity
a. convergent
b. divergent
c. criterion
d. construct
a. convergent
what the test actually predicts demonstrates ___ validity
a. convergent
b. divergent
c. criterion
d. construct
c. criterion
tests that are generally similar but still distinct show ____ validity
a. convergent
b. divergent
c. criterion
d. construct
b. divergent
the trait test is measuring 1 trait
monotrait
data from only one source
monomethod
data from multiple sources
heteromethod
more than one trait is measuring
heterotrait
data is from only one source
monomethod
how well the test predicts behavior/performance is ____ validity
criterion
concurrent criterion validity focuses on
hint: future or current
current
predictive criterion validity focuses on
hint: future or current
future
diagnostic utility, sensitivity — has the disorder, yes or no?
yes, has the disorder
diagnostic utility, specificity — has the disorder, yes or no?
no, does NOT have the disorder
reduce error by increasing consistency
hint: reliability or validity
reliability
variable or characteristic that
influences degree of correlation between 2 other variables
(gender, culture, age, socio-eco background)
moderator variable
gives answer for when and to whom the correlation is significant
hint: reliability or validity
validity
correlation analysis:
r >= 0.70 {what is this for?}
to show variability from true perfect score
correlation analysis:
r >= 0.90 {what is this for?}
hint: kappa coefficient
high stakes testing
what is central tendency?
cling towards the middle of the scale
- leniency/strictness bias
- halo effect/horns effect
- central tendency
these are examples of common biases in ____
interrater reliability
to make r-square change value into a %, what do we do?
move the decimal 2 places