MUED 6440 Ch. 9 Flashcards
Procedure that tests the null hypothesis that the variables in the population correlation matrix are uncorrelated; used for factor analysis with small samples
Bartlett’s sphericity test
Amoutn of variance in each variable accounted for by the factors; equal to the squared multiple correlation of thew variable as predicted from the factors; also equal to the sum of squared loadings for a variable across all factors; provided for each variable
Communalities (hi)
More advanced than explanatory factor analysis; used to test a theory about latent (i.e., underlying, unobservable) processes
Confirmatory factor analysis
Amount of total variance explained by each factor in factor analysis
Eigenvalue
Goal is to describe and summarize data by grouping together variables that are correlated; variables may or may not have been chosen with these underlying structures in mind
Exploratory factor analysis
The process by which the underlying factors from a larger set of variables are determined
Extraction
A mathematical model is created, resulting in the estimation of factors; contrast with principal components analysis
Factor analysis
Provides correlation coefficients between each IV and each factor in the solution; values can also be interpreted as that amount that each IV contributes to each factor
Factor correlation matrix
The Pearson correlations of original variables with factors
Factor loadings
Estimates on the scores participants would have received on each of the factors had they been measured directly
Factor scores
The initial linear combination of IVs; accounts for the largest amount of total variance; equal to the largest eigenvalue for the solution
First principal component
Rotation of factors resulting in factors being correlated with each other and producing several matrices; a factor correlation matrix (i.e., a matrix of correlations between all factors); a loading matrix separated into a structure matrix (i.e., correlations between factors and variables); and a pattern matrix (i.e., unique relationships with no overlapping among factors between each factor and each observed variable) upon which interpretation of factors is obtained
Oblique rotation
Rotation of factors resulting in factors being uncorrelated with each other; result is a loading matrix (i.e., a matrix of correlations between all observed variables and factors) where the size of the loading reflects the extent of the relationship between each observed variable and each factor; interpretation of factors is obtained from the loading matrix
Orthogonal rotation
The result of principal components analysis
Principal components
Most common for extracting factors in factor analysis; original variables are transformed into a new set of linear combinations by extracting the maximum variance from the data set with each component; results in components; contrast with factor analysis
Principal components analysis
