Samenvatting hoorcolleges testtheorie
Inhoud
Week 1: eigenschappen tests & items....................................................................................................4
Hoorcollege 1......................................................................................................................................4
Hoorcollege 2......................................................................................................................................5
Psychologische test.........................................................................................................................5
Meetniveaus...................................................................................................................................5
Variantie.........................................................................................................................................6
Itemscore variantie.........................................................................................................................6
Week 2: verwerken testscores & betrouwbaarheid................................................................................7
Hoorcollege 3......................................................................................................................................7
Bewerkte scores en normen...........................................................................................................7
Vergelijking van normen.................................................................................................................7
Transformaties van standaardscores: lineair...................................................................................7
Transformaties van standaardscores: niet-lineair...........................................................................8
Overzicht van de transformaties.....................................................................................................8
Hoorcollege 4......................................................................................................................................9
[Betrouwbaarheidsinterval]............................................................................................................9
Herhaalbaarheid van metingen......................................................................................................9
Klassieke testtheorie.......................................................................................................................9
Standaardmeetfout.......................................................................................................................10
Week 3: beoordelen van een test (betrouwbaarheid)..........................................................................11
Hoorcollege 5....................................................................................................................................11
Bepaling betrouwbaarheid...........................................................................................................11
Test-hertestmethode....................................................................................................................11
Parallelle vorm methode...............................................................................................................11
Interne consistentie methodes.....................................................................................................11
Nauwkeurigheid van metingen.....................................................................................................14
Vergelijking testscores..................................................................................................................14
Hoorcollege 6....................................................................................................................................15
Kwaliteit van de items (onderlinge samenhang)...........................................................................15
Dimensionaliteit...........................................................................................................................15
Testlengte.....................................................................................................................................15
Heterogeniteit in de populatie......................................................................................................15
1
, Type score die onderzocht wordt..................................................................................................15
Week 4: beoordelen van een test (betrouwbaarheid)..........................................................................17
Hoorcollege 7....................................................................................................................................17
Een goede test ontwikkelen..........................................................................................................17
Item-selectie.................................................................................................................................17
[Correlatie bij dichotome variabelen]...........................................................................................17
Gevolgen van betrouwbaarheid....................................................................................................18
Attenuatie van correlaties.............................................................................................................18
Corrigeren voor attenuatie in X en Y.............................................................................................18
Corrigeren voor attenuatie bij predictieve validiteit.....................................................................18
Week 5: beoordelen van een test (validiteit)........................................................................................20
Hoorcollege 8....................................................................................................................................20
Validiteit........................................................................................................................................20
Vijfhoek van begripsvaliditeit........................................................................................................20
Multi-trek-multi-methode onderzoek...........................................................................................22
Hoorcollege 9....................................................................................................................................23
Een test inzetten in de praktijk......................................................................................................23
Criteriumvaliditeit onderzoeken...................................................................................................23
Dichotome beslissingen................................................................................................................24
Week 6: geavanceerd gebruik tests (IRT)..............................................................................................26
Hoorcollege 10..................................................................................................................................26
Klassieke testtheorie (KTT)............................................................................................................26
Item-responstheorie (IRT).............................................................................................................26
Rasch model.................................................................................................................................26
Birnbaums twee-parameter model (2PLM)..................................................................................27
IRT analyse....................................................................................................................................27
Het doel van IRT............................................................................................................................28
Hoorcollege 11..................................................................................................................................29
Nauwkeurigheid van schattingen..................................................................................................29
Test-informatie functie..................................................................................................................29
Nauwkeurigheid van de meting....................................................................................................29
Testconstructie o.b.v. een itembank.............................................................................................29
Adaptieve tests.............................................................................................................................29
Multi-stage testen.........................................................................................................................30
Klassieke testtheorie vs. item-responstheorie..............................................................................30
Betrouwbare score in IRT..............................................................................................................30
2
,Week 7: test fairness............................................................................................................................32
Hoorcollege 12..................................................................................................................................32
Vormen van testgebruik................................................................................................................32
Bias...............................................................................................................................................32
Construct bias...............................................................................................................................32
Itembias........................................................................................................................................32
Predictieve bias.............................................................................................................................33
3
, Week 1: eigenschappen tests & items
Hoorcollege 1
id X x x2 zx Y y y2 zy xy
1 20 -2 4 -1.41 26 -2 4 -0.71 4
2 22 0 0 0 24 -4 16 -1.41 0
3 21 -1 1 -0.71 28 0 0 0 0
4 24 2 4 1.41 30 2 4 0.72 4
5 23 1 1 0.71 32 4 16 1.41 4
Som 110 0 10 0 140 0 40 0 12
X 22 0 2 0 28 0 8 0 2.4
2
SX 2 2 1 8 8 1
SX 1.41 1.41 1 2.83 2.83 1
N
Gemiddelde = X =
∑ Xi
i=1
N
Deviatiescore = x = X −X
N
Variantie = S2X =
∑ xi2
i=1
N
Standdaarddeviatie = S2 = S 2X √
x
Gestandaardiseerde score = zx =
SX
N
Covariantie = Sxy =
∑ xi y i
i=1
N
S XY
Correlatie = rxy =
S X SY
N
Sum of squares = SS = ∑ x
2
i=1
Variantie-covariantiematrix
X Y
X 2 2.4
Y 2.4 8
Correlatiematrix
X Y
X 1 0.6
Y 0.6 1
4
Inhoud
Week 1: eigenschappen tests & items....................................................................................................4
Hoorcollege 1......................................................................................................................................4
Hoorcollege 2......................................................................................................................................5
Psychologische test.........................................................................................................................5
Meetniveaus...................................................................................................................................5
Variantie.........................................................................................................................................6
Itemscore variantie.........................................................................................................................6
Week 2: verwerken testscores & betrouwbaarheid................................................................................7
Hoorcollege 3......................................................................................................................................7
Bewerkte scores en normen...........................................................................................................7
Vergelijking van normen.................................................................................................................7
Transformaties van standaardscores: lineair...................................................................................7
Transformaties van standaardscores: niet-lineair...........................................................................8
Overzicht van de transformaties.....................................................................................................8
Hoorcollege 4......................................................................................................................................9
[Betrouwbaarheidsinterval]............................................................................................................9
Herhaalbaarheid van metingen......................................................................................................9
Klassieke testtheorie.......................................................................................................................9
Standaardmeetfout.......................................................................................................................10
Week 3: beoordelen van een test (betrouwbaarheid)..........................................................................11
Hoorcollege 5....................................................................................................................................11
Bepaling betrouwbaarheid...........................................................................................................11
Test-hertestmethode....................................................................................................................11
Parallelle vorm methode...............................................................................................................11
Interne consistentie methodes.....................................................................................................11
Nauwkeurigheid van metingen.....................................................................................................14
Vergelijking testscores..................................................................................................................14
Hoorcollege 6....................................................................................................................................15
Kwaliteit van de items (onderlinge samenhang)...........................................................................15
Dimensionaliteit...........................................................................................................................15
Testlengte.....................................................................................................................................15
Heterogeniteit in de populatie......................................................................................................15
1
, Type score die onderzocht wordt..................................................................................................15
Week 4: beoordelen van een test (betrouwbaarheid)..........................................................................17
Hoorcollege 7....................................................................................................................................17
Een goede test ontwikkelen..........................................................................................................17
Item-selectie.................................................................................................................................17
[Correlatie bij dichotome variabelen]...........................................................................................17
Gevolgen van betrouwbaarheid....................................................................................................18
Attenuatie van correlaties.............................................................................................................18
Corrigeren voor attenuatie in X en Y.............................................................................................18
Corrigeren voor attenuatie bij predictieve validiteit.....................................................................18
Week 5: beoordelen van een test (validiteit)........................................................................................20
Hoorcollege 8....................................................................................................................................20
Validiteit........................................................................................................................................20
Vijfhoek van begripsvaliditeit........................................................................................................20
Multi-trek-multi-methode onderzoek...........................................................................................22
Hoorcollege 9....................................................................................................................................23
Een test inzetten in de praktijk......................................................................................................23
Criteriumvaliditeit onderzoeken...................................................................................................23
Dichotome beslissingen................................................................................................................24
Week 6: geavanceerd gebruik tests (IRT)..............................................................................................26
Hoorcollege 10..................................................................................................................................26
Klassieke testtheorie (KTT)............................................................................................................26
Item-responstheorie (IRT).............................................................................................................26
Rasch model.................................................................................................................................26
Birnbaums twee-parameter model (2PLM)..................................................................................27
IRT analyse....................................................................................................................................27
Het doel van IRT............................................................................................................................28
Hoorcollege 11..................................................................................................................................29
Nauwkeurigheid van schattingen..................................................................................................29
Test-informatie functie..................................................................................................................29
Nauwkeurigheid van de meting....................................................................................................29
Testconstructie o.b.v. een itembank.............................................................................................29
Adaptieve tests.............................................................................................................................29
Multi-stage testen.........................................................................................................................30
Klassieke testtheorie vs. item-responstheorie..............................................................................30
Betrouwbare score in IRT..............................................................................................................30
2
,Week 7: test fairness............................................................................................................................32
Hoorcollege 12..................................................................................................................................32
Vormen van testgebruik................................................................................................................32
Bias...............................................................................................................................................32
Construct bias...............................................................................................................................32
Itembias........................................................................................................................................32
Predictieve bias.............................................................................................................................33
3
, Week 1: eigenschappen tests & items
Hoorcollege 1
id X x x2 zx Y y y2 zy xy
1 20 -2 4 -1.41 26 -2 4 -0.71 4
2 22 0 0 0 24 -4 16 -1.41 0
3 21 -1 1 -0.71 28 0 0 0 0
4 24 2 4 1.41 30 2 4 0.72 4
5 23 1 1 0.71 32 4 16 1.41 4
Som 110 0 10 0 140 0 40 0 12
X 22 0 2 0 28 0 8 0 2.4
2
SX 2 2 1 8 8 1
SX 1.41 1.41 1 2.83 2.83 1
N
Gemiddelde = X =
∑ Xi
i=1
N
Deviatiescore = x = X −X
N
Variantie = S2X =
∑ xi2
i=1
N
Standdaarddeviatie = S2 = S 2X √
x
Gestandaardiseerde score = zx =
SX
N
Covariantie = Sxy =
∑ xi y i
i=1
N
S XY
Correlatie = rxy =
S X SY
N
Sum of squares = SS = ∑ x
2
i=1
Variantie-covariantiematrix
X Y
X 2 2.4
Y 2.4 8
Correlatiematrix
X Y
X 1 0.6
Y 0.6 1
4