a method for finding a small set of underlying dimensions from a large set of related measures. The observed data are freely explored in order to discover the underlying (latent) variables that explain the interrelationships among a larger set of observable (manifest) variables. For example, exploratory factor analysis has been conducted to assess whether there are one, two, or more dimensions underlying items used to assess intelligence. The analysis does not yield a unique solution: Although points are fixed in a multidimensional space of underlying factors, the axes according to which one should interpret the factors are not fixed and are infinite in number. These axes may be either
orthogonal (producing uncorrelated factors) or
oblique (producing correlated factors). Compare
confirmatory factor analysis.