1what is validity ?
a criterion from weak , acceptable and good that tells us if the test asseses what is it supposed to assess
2what types of validity are there ?
1. face validity
2. content validiy
3. criterion validity
4. construct validity
3what is face validity ?
valid impression of test on examiners and examinees
no scientific
does it seem relevant
4what is content validity ?
the extent to which the content of a test are representative of the behaviour / construct you want to asses
5by what is charcterised content validity ?
propper selction of items
6why do we use a test matrix ?
optimize content validity of tests
7how can we optimize content validity ?
1. which items are to be measured ?
2. item property
3. goal of the test
4. panel of experts
8why do we use discrepnacy of which items are to be measured ?
need previous knowledge and consesus of the participant
9what are item properties ?
how difficult the items are
how are they answered ( yes or no )
10what are possible goals of tests ?
relative tests compare to norm
absolute test compared to previously determined criterion
11how do we use panel of experts ?
they have the expirince to determine the relevance of items
12how does a panel of expert table look like ?
13what is the criterion validity ?
the correlation between test scores and behaviour / construct you want to predict
1. concurrent validiy
2. predictive validity
the relation between test score X and criterion Y
14what are the two subtypes of criterion validity ?
1. concurrent validity
2. predictive validity
15what is concurent validity ?
bahaviour at same time
16what is predictive validity ?
bevaiour now vs in the future
17what requirments are there for criterion validity ?
high reliability
high relevance
free from contamination (no overlap )
18what difficulties occur in criterion validity ?
difficult to find a good criterion measure
criterion is often a test as well
criterion is influenced by more factors
19why is the reliabaility important for the criterion validity ?
max criterion validity of a test is determined by reliability of the test and of the criterion
20what levels of reliability correspund to which criterion validity ?
test reliability .90 = criterion .90
test reliability .10 = criterion .30
test reliability .50 = criterion .39
21what is the standard error of estimate in criterion validity ?
the margin of error you can expect in the predicted criterion score
22what is the SE este formula ?
23what do each of the components in the SE est mean ?
SD y = standard devioatoon of criterion score Y
rxy = correlation between x and y
24how do we gain a ****predictive validity?
1. determine criterion score Yi using test score Xi
2. calculate the standard error estimate with the formula
3. determine confidence interval with the CI =Yi +_90/95/99% x SEest
*sometimes we use the regression equation
25how to use the regression equation ?
the same as Se estimate but we use a formula given
use the score of the participant instead of X in the formula
26what is the decision theory?
estimates error by :
1. hits
2. false positive
3. false negative
27decision theory acceptable levels of validity?
aleweys estimation error :
1. hits
2. false positive
3. false negative
1. sensitivty
2. specificity
3. test cut off
4. prevelence
28what are hits ?
proportion of correct
29how do we clacuate hits ?
yey:ye + no:no / total
30what are false positivs ?
prooportions falsly diagnosed
31how can we calcute false positivs ?
yes:no / total
32what are false negatives ?
proportions flasely not diagnosed
33how can we calculate false negatives ?
no:yes / total
34what is sensitivity ?
proportions that have diesese and are correctly classified
35what is specificity ?
proportion of healthy people corrctly classified
36how can we calculate sensitivity ?
yes:yes / yes:yes + yes:no
37how can we calculate specificty ?
no:no / yes:no + no:no
38what are test cut offs ?
overall proportion classified by test
39what are prevalance ?
how often the diesese occur in reality ?
40how can we calculate test cut offs?
yes:yes +yes:no / total
41how can we calculate prevelance ?
yes:yes + yes:no / total
42what is construct validity ?
the extent to which the test score is a good reflection of the construct you want to asses
43what is a construct ?
variable that cannot be observed or measured directly
based on theoretical assumptions
44what is an exmaple of a construct ?
intelligence
45what are the ways of finding construct validity?
1. relation with other tests
2. factor analysis analysis of components
3. compare groups group differences and intervention effects
46what is the **relation with othes tests**?
> relationship between test scores and scores on other variables
1. convergence validity measure the same construct (need high )
2. dicriminant validity measures another construct ( need low)
→correlate the test score with the score of same / other tests
47what are the two subtypes of construct validity ?
1. convergence validity
2. discriminant validity
48what is cinvergent validity ?
measure the same construct ( need to be low )
49what is discriminant validity ?
extent to whcih they measure another construct ( need to be low)
50what is the difference between convergent and concurrent validity ?
convergent correspondent between your instrument and another instrument
concurrent predict the validity at the same time assesed
51what is the multitrait ?
extent to which the score relates to scores on test assessing other traits
discriminant validity
52what is the multimethod multitrait matrix ? (MMM)
simultanious assesment of convergence and discriminant validity and reliability
53what is the multimethod ?
using different methods to determine the same behaviour
convergent validity
54how do we need out values to be ?
high A > C > B > D low
55what are the components of the matrix ?
a reliability as its the same
c convergent validity
discriminant :
b different construct same methods
d different construct different methods
56what is factor analysis ?
method to examine independent constructs that a test contains
57how do we analyse the factors ?
identify as few factors that explain as much of covariance
1. Exploratory factor analysis (EFA) NO SPECIFIC EXPECTATIONS
2. Confirmatory factor analyse (CFA)
58how do you compare groups ?
same test different groups or at different times
59what means when a test is not valid ?
not reliable
not valid
theory based on no correct
test does not measure what you wnat to asses
test measures something you do not want
60what to do when not valid ?
adjust cut off score
more strict / less strict
61what are other threaths ?
test results
does it benefit to diagnose ?
validity shrinkage( cross validation )