Lecture notes Measurement Theory & Assessment I

Same rate for the variante (02)
Classical test theory ( CTT ) : X = T te
→
0×2=0+2+0,2
→
ob served score =
Erve score t Error p p î


The Ieliabitity is the ratio ( verhouding ) between the Observer score (X) and
the truc score ( T )


trueteamiiir
de waarheid '
IImetselaar
←
↳±
0×2


The smaller the error ( e) ,
the bette the ratio between T and X
→
hogere reliabitity
68% CI → Xi I 1 '

SDe
90 % CI → Xi ± 1,65 .

5De
95 % CI → Xi ± 1,96
-

5 De → This on e is used mast comm only
99 % CI → Xi ± 3 .
5 De
↳ individual score




the test reliable
Is erough ?
Important decisions : less important decisions :
Group level :

-

Goods 90 .
-
Goed > .
80 -

Goods .
70
-

Sufficient > 80 90 .
-

.
-


Sufficient > .
80 -

.
70 -

Sufficient > 70 60 .
-

.




-
In sufficient c. 80 -
In sufficient c. 70 -
In sufficient c. 60




Percepties Corre spon
ding to Standard device ons :




>
2 =
% individuals with score c- 2513
316 =
% individuals with score c- 1 SD




Éne
P 50 % individuals with score <
score
average
=




384 =
% individuals with score et 1 SD
> 98 = % individuals with score c -125 D




( Youaangifte
mbar these threeand
, III soon
"
-
The distrance between P 16 )
c -
1 SD
is the difference between P 84 (c +1 SD ) and P 50 (0 SD ) =
34
-
The distrance between P2 (c -
2513 ) and 750 (0 SD ) = 48
,
50

is the difference between P 98 (c +2513 ) and P 50 (0 SD ) = 48




Minimum sample sizes for :


N
-



Important decisions ( irreversibele → >
300
-
Less important decisions → n 200 >


-


Group level → N 100 >