1.5

the iq test was not valid for the purpose of classifying workers

evidence that it has validity for the particular stiauaion to which it is being applied

reason

the association between the results of a test and other variables

it can be defined as the agreement between a test score or measure and the quality it is belived to measure

-construct-related validity

-criterion-related validity

-content-related validity

the test appears to measure what it says its measuring

no

it can help motivate test takers bc thet can see that the items are relevant to the skill or quality that is thought to be measured

educational testing

the content of the material we covered in class

logical

not all important aspects f the contruct are sampled

ex) SAT only had algebra and not geometry

scores are influenced by factors irrelevant to the contruct

ex) anxiety or reading problems

how well a test corresponds with a particular criterion

*correlation between a test and a well defined criterion measure

using a test to forecast something

*SAT is the predictor variable and college GPA is criterion

when one measure is used to estimate another event taking place at the same time

*simultaneous relationship between the test and the criterion

compares peoples interests with those of individuals who are in a variaety of occupations and are happy in them

is a correlation expressing a relationship between a test and a criterion

.60

.3 to .4

reliability

when there are no other reliable predictors available

• is a rating of a patient by a professional
• *nurse or doctor

is a self reporting scale of depression

SAT-II

the logic of criterion validation presumes that the conditions in effect when the test was validated will still be in effect when the new results are applied

the criterion itself must be reliable and valid

the population on which the test was validated may not be the same as the population on which it will be used on.

ex)if the test was developed on white people, it may not be the same for hispanics

if a validity coefficient was based on a very small sample the results can be quite misleading

ex) a student was admitted even though he did not meet the standard GRE score. He was not allowed to graduate till he met the standard GRE score

ex) the person doesnt get in to ivy league law school unless your scores are really high. this creates a restricted range

ex) can findings obtained in one situation actually be generalized to other situations?

predictive relationships may not be the same for all groups

ex)when a test is used to measure indifcudals who do are not native englsih speakers, the validity of the test may be much different

the most accurate measure of the phenomenon being observed

a researcher simultaneously defines some construct and then develops instrumentation to measure it

-attachment

-caring

-intimacy

when a measure correlates well with other things thought to be measures of a certain construct

ex) if we developed a health index then people who score low on it shold have more physical symptoms than those taht score high on it

your test should not correlate with measures of unrelated constructs

ex) if your measure correlates too highly with other measures of your construct, why doe it have any purpose at all? why not just use one?

is a test that wants to predict to a particular criterion.

Ex) achieving a higher score than 600 on GRE, predicting if a student with a certian score will do better than 75 percent of classmates

is a test where you look at where this score falls in the population being measured

near zero

very high bc so many other factors can affect a variable between the test and the actual event

contruct validity

no, its impossible

it correlates with itself

• square root of the product of the reliability of the test and the reliability of the criterion
•  ex) IQ scores and school grades