Lecture 2.1 Flashcards
1
Q
reliability score
A
- always between 0 and 1
- correlation between scores
- e.g. items, trials
- reflects repeatability and consistency
= ration actual vs test score
2
Q
error score and T
A
- positive: T larger than X
- negative: T smaller than X
3
Q
causes of measurement error
A
- item selection
- sample (also of behavior)
- examinee
- examiner
4
Q
systematic errors
A
- via manual or test construction
- validity
5
Q
unsystematic errors assumptions
A
- random
- average to 0 when data is large enough
- not related to eachother
- are not realted to T
- affect consistency
- reliability
6
Q
Three assumptions about CTT
A
- focus on unsystematic errors
- are normally distributed
- CI is most accurate
7
Q
within one SD
A
- 67.2%
8
Q
within two SD
A
- 95.4%
9
Q
reliability formula
A
- true score diveded by observed score
10
Q
small CI
A
- instrument is reliable
11
Q
kinds of CIs
A
- 68
- 95
- 98
12
Q
methods of reliability
A
- test-retest
- alternate forms
- Spearman Brown/split half
- coefficient alpha
- KR20
- inter-scorer / inter-rater
13
Q
disadvantages of test-retest
A
- practice effects
- prone to fluctuation in construct
14
Q
alternate forms advantages
A
- no practice effects
15
Q
alternate forms disadvantages
A
- version might measure something different
- very expensive
