1.4.1 psych test
Card Set Information
1.4.1 psych test
In Psych test, what does error imply?
it means that there will always be some inaccuracy in our measurements
Test that are relatively free of measurement error are considered?
Psychological tests deal with complex traits that are...
*intelligence and aggresiveness
What did Charles Spearman do?
worked out most of the basics of contemporary reliability theory
*published it in 1904 in American journal of psychology
What book did Thorndlike publish?
an introduction to the theory of mental and social measurements
basics of test score thoery- What is an observed score (x) composed off? (2)
-true score (T)
basics of test score thoery- What shape is the distribution of random errors?
normal or bell shaped
basics of test score thoery- In a very accurate test the distribution is ____
basics of test score theory- In a very innacurate test the distribution can be very..
What is the standard error of measuremnt?
is the SD of the error scores around the true score
what 2 things are used to estimate the standard error of measurement?
-standard deviation of the observed scores
-reliability of the test
Because measuring instruments are imperfect, the score observed for each person..
almost always differs from th persons true ability or characteristic
the difference between the true score and the observed score is..
the measurement error
What does the item response theory do to the tests? (3)
-it shortens the test
-it is believed to provide a more reliable estimate of the individuals true ability
what does somebody have to do if they want to use a test to make an employment or educational plaement decisions?
-they need to have it approved by federal govt
*fed govt need to see that it is reliable
What does a score of reliability .40 mean? (2)
-40% of the difference among the people tested with it can be predicted from the test results
-60% must be ascribed to random or chance factors
What is a big part of psych testing?
measuring differences between individuals
what are 3 ways test reliability can be estimated?
-parallel forms method
-internal consistency method
3 ways to test reliability- which two ways are basically similar?
3 ways to test reliability-how many times is the test-retest method administered?
2 different times
3 ways to test reliability- what does the test-retest method used for?
applies when measuring stable traits
3 ways to test reliability- What can happen when adminitering a test-retest method exam?
3 ways to test reliability- What is test-retest method an example off?
intelligence and achievement tests
3 ways to test reliability-test-retest method- What does a low correlation between two scores on 2 administrations mean?
the characteristic is not stable
3 ways to test reliability- What is the parallel forms method like?
tests the person on 2 different forms of the same test
*the items should measure the same construct but the items should be different
3 ways to test reliability- which provides one of the rigorous assesments of relaibility?
parallel forms method
3 ways to test reliability- decription of split-half method
-divides the items in half and seeing if you get the same basic socres on one half of the items as you do on the other
*items that are selected for each half should be good samples of the items on the whole test
3 ways to test reliability- What is anohter way to administer a split-half method?
Methods for estimating the internal consistency of a test- KR20
estimates internal consistency by simultaneously considering all possible ways of splitting the items
Methods for estimating the internal consistency of a test- What is the coefficient alpha?
it estimates the internal consistency item by item of tests in which the items are NOT scored 1 or 0
Methods for estimating the internal consistency of a test- who made the coefficient alpha?
What do all measures of internal consistency evaluate?
the extend to which the different items on the test measure the same ability or trait
Difference score on reliability
Sometimes the difference score is what we are trying to acquire
*a person coming after treatment or after training
For what method is the kappa statisitc useful in?
it is the best method for estimating interrater reliability
What is direct observations of behavior?
a behaviorally-oriented psychologists would record the number of instances that a certain behavior occurs
What does the kappa stat indicate?
the actual agreement as a proportion of potential agreement following a correction for chance agreement
What does the standard error of measuremnt estimate?
the degree to which a test provides innacurate
How many SDs lower are the students from non college preparatory scoring thn college prep students?
What reliability score would be good enough for resesarch purposes?
.7 to .8
Would a reliability of .9 be good enough for when tests are used to make important decisions about someones future?
no, it should be higher
what to do with low reliability tests- what would increasing the number of items do for a test?
increase the reliability
what to do with low reliability tests- How can cronbach's alpha help out in this?
it can point out items that do not actually measure the construct,leaving them in reduces the tests reliability
what to do with low reliability tests- what does the correction for attentuation do?
it estimates what the correlation between tests would have been if there had been no error of measurement
There can be no validity without...
If a test is not reliable then
one cannot demonstrate that it has any meaning