lecture 4 *diagnostics *

39 flashcardsby M
Study these flashcardsFree · spaced repetition · iPhone, Android, Mac & web

All cards

1
what is the standard error of measurments ?
a measure of how much measured test scores are spread around a "true" score
2
what **assumptions** does SEM have ?
1. varince of measurment error = all individuals taking test
2. measurment error is normally distributed
1. mean = M = o , standard deviation = SDe
3. standard deviation of measurment error is a measure of the expected deviation of X relative to T
3
what is the SEM formula ?
4
what does SEM stand for ?
SEM = standard error of measurment
5
what does SDx stand for ?
SDx = standard deviation of X in sample
6
what does rxx stand for ?
rxx = (estimated) reliability of the test
7
what is SEM in perfectly reliable test ?
r will be 1
8
whi r= 1 is a perfectly reliable test ?
squareroot of 0 = 0

SEM = 0
9
what is a CI ?
CI = intterval in which with a certain percentage of certainty lies the individual's true score
10
what range score for CI are there ?
range of scores within which true score lies

68 90 95 99%
11
how do we get from SEM to CI?
1. administer test and determine score
2. calulate standard measuremnt error SDe of the test
3. claculate CI
12
how to conduct **step 1**?
complete questionnaire by participant or parent
we need the score itself
13
how to conduct **step 2**?
use the formula to calculate
we need reliability rxx of test
we need SDx of the test
14
what is the 68 % CI interval ?
Xi +/ _ 1 x SDe
15
what is the 90 % CI interval ?
Xi + 1.65 x SDe
16
what is the 95 % CI interval ?
**Xi + 1.96 x SDe**
17
what is the 95 % CI interval ?
Xi + 3 x SDe
18
how do we conduct step 3?
calculate the lower limit and upper limit of the score
19
what confidence interval is usually used?
95%

lower limit : Xi _ 1.96 x SDe

upper limit : Xi + 1.96 x SDe
20
whan do we use SE diff ?
when we want to find the difference in 2 scores
21
what do we need in order to find SEdiff?
M and SD
22
what do we use in order to calculate difference?
SE diff
23
what are the formula used ?
24
how do we use SEdiff to calcualte CI for difference scores?
1. determine difference between scores
2. calculate SEdiff
3. calculate CI 95% of difference score
25
how do we determine difference between scores ?
initial _ actual
26
how do we calculate SE diff ?
1. use reliability rxx of both tests and SDx of both tests
2. use measurment error SDe of both tests
27
how do we calcualte CI 95% of difference score ?
28
when is there a difference ?
if the CI does not include zero , there is a real difference
29
what could be another approach ?
calcualte confidence interval around 0
why ? know which range difference scores could actually be 0
30
what are the steps of calculating the CI around 0 ?
1. no difference , use 0
2. calcualte SEdiff (with formula)
3. calualte CI interval
31
when is there no actual difference if we calulate CI around 0 ?
*scores in that range imply no actual difference
32
how do we calcualte differecne if we test with different instrument?
1. chack the difference between scores
2. use the formula based on what info we have
3. calculate CI
4. coclude : does it include 0?
33
how do we find an acceptable level of reliability in terms of context?
what is the "price " of a wrong decision ?

how much on an influence does the test have ?
34
how do we find the acceptable level of reliability in terms of level ?
individual : needs higher reliability

groups / reserch : requires less reliability
35
what is the interval of reliability for an important decision ?
Good : more than .90

sufficinet : between .80 and .90

insufficint : less than 80
36
what is the interval of reliability for a less important decision ?
good : more than .80

sufficint : between .70 and .80

insufficint : less than .70
37
what is the interval of reliability for a group level decision ?
good : less than .70

sufficint : between .60 and .70

insufficint : less than .60
38
what alternatives of CTT are there ?
IRT : item response theroy
39
what is IRT ?
measurment error depends on test score

reliability items may be distinguished between individuals

charachterisitcs of items

1. difficulty
2. discrimination index
3. measure fir unformative function

we test the "test construcution "
Study all 39 cards for free