| Factor Analyses (FA) are a set of techniques that allows us to look in more detail at the strength of underlying correlations between scores on the items of a measure. These provide some useful indication about whether items appear to tap common, i.e. shared, sources of variation between individuals, a major part of construct validity. Historically FA was exploratory: Exploratory Factor Analysis (EFA), actually often a closely related approach called Principal Component Analysis (PCA) was often used sometimes under the name EFA though it’s actually a distinct method. Less often, Multi-Dimensional Scaling (MDS) has been used for similar purposes but it only analyses ordinal correlations not correlations between raw item scores. More recently, with much more computer power avialable, Confirmatory Factor Analysis (CFA), which can test an expected mapping of items to presumed factors has become fashionable. |
Detail #
See Confirmatory Factor Analysis (CFA) and Exploratory Factor Analysis (EFA) for more detail, Construct Validity for background detail. Principal Component Analysis (PCA) and Multi-Dimensional Scaling (MDS) are background, not central to FA.
Try also … #
Bifactor analysis
Confirmatory Factor Analysis (CFA)
Construct Validity
Exploratory Factor Analysis (EFA)
Multi-Dimensional Scaling (MDS)
Principal Component Analysis (PCA)
Chapters #
Not developed in the OMbook.
Further reading #
Should be developed in the Rblog but not there yet.
Online applications #
Nothing likely.
Dates #
Started before 14.xi.21, last tweaked 22.i.25.
