Lecture 2.2 Flashcards
1
Q
assummptions for SEM
A
- variance same for all s
- normally distributed
SEM SD is a measure of deviation X from T
2
Q
SEM score impliaction
A
- 0 = perfectly reliable
- dependent on scale and scoring range, is relative
3
Q
CI usage
A
- with SD e
- always rounded to possible numbers WITHIN CI
4
Q
SEM dif
A
- difference score
- two formulas
- should have approx. same mean and SD
5
Q
SE dif
A
- takes into account unreliability of both tests
6
Q
accepted reliability depends on
A
- price and influence of decision
- how many tests are used
- about group or individual
7
Q
good reliability
A
- 90 and more
8
Q
sufficient reliability
A
- between 80 and 90
9
Q
insufficient reliability
A
- below 80
10
Q
Item Response Theory
A
- alternative to CTT
- more modern
- measurement error not the same for every s
- higher performance means less error
- item difficulty and discrimination index = reliability of each item
