in
factor analysis, the repositioning of
latent variables (factors) to a new, more interpretable configuration by a set of mathematically specifiable
transformations. Factors initially are extracted to meet a mathematical criterion of maximal variance explanation, which often does not result in a scientifically meaningful representation of the data. Indeed, for any one factor solution that fits the data to a specific degree, there will exist an infinite number of equally good mathematical solutions. Thus, rotation is required to obtain a solution that is both mathematically viable and logically sound. See
oblique rotation;
orthogonal rotation.