Factor analysis:
Type
- Independent
- Multivarivte
Fvctoar vnvlysis : Confarmvtoary vnvlysis
Exploarvtoary vnvlysis: R type = garoups difearent ivarivbles into fvctoars
Q type = garoups difearent cvses into clustears
Input & sample
- Intearivl/arvto
- Semvntc/Likeart scvle
- 3-5 items pear fvctoar
- Pvarsimonious (economicvl)
- Gvarbvge in – gvarbvge out: only areleivnt items
- Min 100 arespondents
- 10:1 arvto
- High cvse-to-ivarivble arvto: no oiearfing
Objectie
- Dvtv areducton
- Find vn undearlying dimension in the dvtv
Type explanaton
- Exploarvtoary: parincipvl component vnvlysis.
The fvctoar is v linevar combinvton of stvndvardized oariginvl scoares
F1 = A1X1 + A2X2 + … + ANXN
The oariginvl scoares hvie v weight =>
Fvctoar = weighted index of oariginvl scoares (foar exvmple; consumear parices index)
- Confarmvtoary vnvlysis: tarue fvctoar vnvlysis.
The obsearied ivarivble (x1) mevsuares pvart of the undearlying lvtent ivarivble (F1) + eararoar.
A lvtent ivarivble cvn’t be mevsuared by just one queston(, foar exvmple pearsonvlity). You hvie to vsk
difearent questons in oardear to fnd the pearsonvlity.
X1 = A1F1 + E1
X2 = A2F2 + E2
Eieary obsearied ivarivble hvs v weight, this weight shows how well it areparesents the lvtent ivarivble.
So the oariginvl scoare = lovding on v fvctoar + eararoar
Factor extracton metthod
- Totvl ivarivnce is common ivarivnce
- Vvarivnce between 2 objects = squvared coararelvton
Diferences:
Totvl ivarivnce => parincipvl component vnvlysis:
It uses the totvl ivarivnce & it is to areduce the dvtv
Common ivarivnce => common fvctoar vnvlysis:
It uses the common ivarivnce & it is used to fnd vn undearlying dimension (lvtent
dimensionvlity)
Type
- Independent
- Multivarivte
Fvctoar vnvlysis : Confarmvtoary vnvlysis
Exploarvtoary vnvlysis: R type = garoups difearent ivarivbles into fvctoars
Q type = garoups difearent cvses into clustears
Input & sample
- Intearivl/arvto
- Semvntc/Likeart scvle
- 3-5 items pear fvctoar
- Pvarsimonious (economicvl)
- Gvarbvge in – gvarbvge out: only areleivnt items
- Min 100 arespondents
- 10:1 arvto
- High cvse-to-ivarivble arvto: no oiearfing
Objectie
- Dvtv areducton
- Find vn undearlying dimension in the dvtv
Type explanaton
- Exploarvtoary: parincipvl component vnvlysis.
The fvctoar is v linevar combinvton of stvndvardized oariginvl scoares
F1 = A1X1 + A2X2 + … + ANXN
The oariginvl scoares hvie v weight =>
Fvctoar = weighted index of oariginvl scoares (foar exvmple; consumear parices index)
- Confarmvtoary vnvlysis: tarue fvctoar vnvlysis.
The obsearied ivarivble (x1) mevsuares pvart of the undearlying lvtent ivarivble (F1) + eararoar.
A lvtent ivarivble cvn’t be mevsuared by just one queston(, foar exvmple pearsonvlity). You hvie to vsk
difearent questons in oardear to fnd the pearsonvlity.
X1 = A1F1 + E1
X2 = A2F2 + E2
Eieary obsearied ivarivble hvs v weight, this weight shows how well it areparesents the lvtent ivarivble.
So the oariginvl scoare = lovding on v fvctoar + eararoar
Factor extracton metthod
- Totvl ivarivnce is common ivarivnce
- Vvarivnce between 2 objects = squvared coararelvton
Diferences:
Totvl ivarivnce => parincipvl component vnvlysis:
It uses the totvl ivarivnce & it is to areduce the dvtv
Common ivarivnce => common fvctoar vnvlysis:
It uses the common ivarivnce & it is used to fnd vn undearlying dimension (lvtent
dimensionvlity)
