Exploratory factor analysis estimates latent dimensions that account for shared covariance among observed variables. Factor number, extraction, rotation and interpretability require explicit judgement.
Read the full entry →Research Wiki · Methods Lab
Factor and dimension-reduction techniques
Methods that compress correlated variables into lower-dimensional structures while retaining interpretable information.
Methods family
Methods that compress correlated variables into lower-dimensional structures while retaining interpretable information.
Use methods only after defining the estimand, data structure, assumptions, validation plan and decision consequence.
Confirmatory factor analysis tests a pre-specified measurement structure linking observed indicators to latent factors. Fit, reliability, validity and alternative models should be considered together.
Read the full entry →Principal component analysis transforms correlated numeric variables into orthogonal components that capture descending amounts of variance. It is a data-reduction method, not automatically a latent-construct model.
Read the full entry →Common-factor analysis is a method within factor and dimension-reduction techniques. Methods that compress correlated variables into lower-dimensional structures while retaining interpretable information. Its usefulness depends on data structure, assumptions, validation and whether the output answers the intended decision.
Read the full entry →Principal-axis factoring is a statistical or analytical concept within factor and dimension-reduction techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Read the full entry →Maximum-likelihood factor analysis is a method within factor and dimension-reduction techniques. Methods that compress correlated variables into lower-dimensional structures while retaining interpretable information. Its usefulness depends on data structure, assumptions, validation and whether the output answers the intended decision.
Read the full entry →Image factoring is a statistical or analytical concept within factor and dimension-reduction techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Read the full entry →Alpha factoring is a statistical or analytical concept within factor and dimension-reduction techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Read the full entry →Q-factor analysis is a method within factor and dimension-reduction techniques. Methods that compress correlated variables into lower-dimensional structures while retaining interpretable information. Its usefulness depends on data structure, assumptions, validation and whether the output answers the intended decision.
R-factor analysis is a method within factor and dimension-reduction techniques. Methods that compress correlated variables into lower-dimensional structures while retaining interpretable information. Its usefulness depends on data structure, assumptions, validation and whether the output answers the intended decision.
Bifactor analysis is a method within factor and dimension-reduction techniques. Methods that compress correlated variables into lower-dimensional structures while retaining interpretable information. Its usefulness depends on data structure, assumptions, validation and whether the output answers the intended decision.
Second-order factor analysis is a method within factor and dimension-reduction techniques. Methods that compress correlated variables into lower-dimensional structures while retaining interpretable information. Its usefulness depends on data structure, assumptions, validation and whether the output answers the intended decision.
Higher-order factor analysis is a method within factor and dimension-reduction techniques. Methods that compress correlated variables into lower-dimensional structures while retaining interpretable information. Its usefulness depends on data structure, assumptions, validation and whether the output answers the intended decision.
Dynamic factor analysis is a method within factor and dimension-reduction techniques. Methods that compress correlated variables into lower-dimensional structures while retaining interpretable information. Its usefulness depends on data structure, assumptions, validation and whether the output answers the intended decision.
Bayesian factor analysis is a method within factor and dimension-reduction techniques. Methods that compress correlated variables into lower-dimensional structures while retaining interpretable information. Its usefulness depends on data structure, assumptions, validation and whether the output answers the intended decision.
Sparse factor analysis is a method within factor and dimension-reduction techniques. Methods that compress correlated variables into lower-dimensional structures while retaining interpretable information. Its usefulness depends on data structure, assumptions, validation and whether the output answers the intended decision.
Non-negative matrix factorisation is a statistical or analytical concept within factor and dimension-reduction techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Read the full entry →Independent component analysis is a method within factor and dimension-reduction techniques. Methods that compress correlated variables into lower-dimensional structures while retaining interpretable information. Its usefulness depends on data structure, assumptions, validation and whether the output answers the intended decision.
Correspondence analysis maps associations in a contingency table by decomposing chi-square distances into dimensions. The map is descriptive and must be interpreted with contributions and inertia, not visual proximity alone.
Read the full entry →Optimal scaling is a statistical or analytical concept within factor and dimension-reduction techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Homogeneity analysis is a method within factor and dimension-reduction techniques. Methods that compress correlated variables into lower-dimensional structures while retaining interpretable information. Its usefulness depends on data structure, assumptions, validation and whether the output answers the intended decision.
Multidimensional scaling is a statistical or analytical concept within factor and dimension-reduction techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Metric multidimensional scaling is a statistical or analytical concept within factor and dimension-reduction techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Non-metric multidimensional scaling is a statistical or analytical concept within factor and dimension-reduction techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
t-SNE is a statistical or analytical concept within factor and dimension-reduction techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
UMAP is a statistical or analytical concept within factor and dimension-reduction techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Autoencoder-based dimension reduction is a statistical or analytical concept within factor and dimension-reduction techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Browse linked method entries
Exploratory factor analysis
Exploratory factor analysis estimates latent dimensions that account for shared covariance among observed variables. Factor number, extraction, rotation and interpretability require explicit judgement.
Open methodConfirmatory factor analysis
Confirmatory factor analysis tests a pre-specified measurement structure linking observed indicators to latent factors. Fit, reliability, validity and alternative models should be considered together.
Open methodPrincipal component analysis
Principal component analysis transforms correlated numeric variables into orthogonal components that capture descending amounts of variance. It is a data-reduction method, not automatically a latent-construct model.
Open methodCommon-factor analysis
Common-factor analysis is a method within factor and dimension-reduction techniques. Methods that compress correlated variables into lower-dimensional structures while retaining interpretable information. Its usefulness depends on data structure, assumptions, validation and whether the output answers the intended decision.
Open methodPrincipal-axis factoring
Principal-axis factoring is a statistical or analytical concept within factor and dimension-reduction techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodMaximum-likelihood factor analysis
Maximum-likelihood factor analysis is a method within factor and dimension-reduction techniques. Methods that compress correlated variables into lower-dimensional structures while retaining interpretable information. Its usefulness depends on data structure, assumptions, validation and whether the output answers the intended decision.
Open methodImage factoring
Image factoring is a statistical or analytical concept within factor and dimension-reduction techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodAlpha factoring
Alpha factoring is a statistical or analytical concept within factor and dimension-reduction techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodQ-factor analysis
Q-factor analysis is a method within factor and dimension-reduction techniques. Methods that compress correlated variables into lower-dimensional structures while retaining interpretable information. Its usefulness depends on data structure, assumptions, validation and whether the output answers the intended decision.
Open methodR-factor analysis
R-factor analysis is a method within factor and dimension-reduction techniques. Methods that compress correlated variables into lower-dimensional structures while retaining interpretable information. Its usefulness depends on data structure, assumptions, validation and whether the output answers the intended decision.
Open methodBifactor analysis
Bifactor analysis is a method within factor and dimension-reduction techniques. Methods that compress correlated variables into lower-dimensional structures while retaining interpretable information. Its usefulness depends on data structure, assumptions, validation and whether the output answers the intended decision.
Open methodSecond-order factor analysis
Second-order factor analysis is a method within factor and dimension-reduction techniques. Methods that compress correlated variables into lower-dimensional structures while retaining interpretable information. Its usefulness depends on data structure, assumptions, validation and whether the output answers the intended decision.
Open methodHigher-order factor analysis
Higher-order factor analysis is a method within factor and dimension-reduction techniques. Methods that compress correlated variables into lower-dimensional structures while retaining interpretable information. Its usefulness depends on data structure, assumptions, validation and whether the output answers the intended decision.
Open methodDynamic factor analysis
Dynamic factor analysis is a method within factor and dimension-reduction techniques. Methods that compress correlated variables into lower-dimensional structures while retaining interpretable information. Its usefulness depends on data structure, assumptions, validation and whether the output answers the intended decision.
Open methodBayesian factor analysis
Bayesian factor analysis is a method within factor and dimension-reduction techniques. Methods that compress correlated variables into lower-dimensional structures while retaining interpretable information. Its usefulness depends on data structure, assumptions, validation and whether the output answers the intended decision.
Open methodSparse factor analysis
Sparse factor analysis is a method within factor and dimension-reduction techniques. Methods that compress correlated variables into lower-dimensional structures while retaining interpretable information. Its usefulness depends on data structure, assumptions, validation and whether the output answers the intended decision.
Open methodNon-negative matrix factorisation
Non-negative matrix factorisation is a statistical or analytical concept within factor and dimension-reduction techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodIndependent component analysis
Independent component analysis is a method within factor and dimension-reduction techniques. Methods that compress correlated variables into lower-dimensional structures while retaining interpretable information. Its usefulness depends on data structure, assumptions, validation and whether the output answers the intended decision.
Open methodMultiple correspondence analysis
Correspondence analysis maps associations in a contingency table by decomposing chi-square distances into dimensions. The map is descriptive and must be interpreted with contributions and inertia, not visual proximity alone.
Open methodOptimal scaling
Optimal scaling is a statistical or analytical concept within factor and dimension-reduction techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodHomogeneity analysis
Homogeneity analysis is a method within factor and dimension-reduction techniques. Methods that compress correlated variables into lower-dimensional structures while retaining interpretable information. Its usefulness depends on data structure, assumptions, validation and whether the output answers the intended decision.
Open methodMultidimensional scaling
Multidimensional scaling is a statistical or analytical concept within factor and dimension-reduction techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodMetric multidimensional scaling
Metric multidimensional scaling is a statistical or analytical concept within factor and dimension-reduction techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodNon-metric multidimensional scaling
Non-metric multidimensional scaling is a statistical or analytical concept within factor and dimension-reduction techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodt-SNE
t-SNE is a statistical or analytical concept within factor and dimension-reduction techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodUMAP
UMAP is a statistical or analytical concept within factor and dimension-reduction techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodAutoencoder-based dimension reduction
Autoencoder-based dimension reduction is a statistical or analytical concept within factor and dimension-reduction techniques. It should be selected for the data-generating process and decision question rather than because software makes it available.
Open methodNo entries match this search.