Karteikarten
What to do when running an FA
Karten in diesem Stapel (9)
- 11) Look at the correlation matrix and check for...Items too highly correlated = >.8
Determinant >.00001= no multicollinearity
items with low correlation <.3= not good - 22) Look at KMO and Bartlett's testKMO > 0.5 = good
Bartlett <0.05 = good - 33) Look at the total variance explainedSee what SPSS has extracted and how much variance they explain individually and as a cumulative figure
Check how many factors have greater eigenvalues than 1. - 44) Look at the Scree plotSee if you agree with SPSS.
- 55) Look at the pattern matrixAssign each item into a factor and then name that factor.
- 66) Look at factor correlation matrix (if done an oblique rotation)An see if it's justified
- 7What is a rotation?It is maximising the loading of questions onto factors
- 8Orthogonal rotation?Factors remain maximally unrelated to each other.
Look at Rotated Factor Matrix table
E.g. Varimax rotation - 9Oblique rotation?Assumes factors can be correlated- enables rotation to fit both factors.
Look at Pattern matrix table and check factor correlation to see if justified- ignore structure.
E.g. Direct Oblimin
Informationen
- Beschreibung
- Karteikarten am What to do when running an FA, erstellt von franceskeel am 27/01/2015.
- Erstellt von
- franceskeel
- Aufrufe
- 1
- Bewertung