Varimax Rotation

An orthogonal factor rotation maximising the variance of squared loadings within each factor, tending to make each variable load strongly on only one factor.

Source: Kaiser 1958

Varimax rotation is an orthogonal rotation that reworks a factor solution to make each factor as distinct as possible, by maximising the spread of the squared loadings within each factor. This pushes every variable toward loading strongly on just one factor and weakly on the rest, giving a clean structure that is easy to interpret while keeping the factors uncorrelated. It is a common default when the underlying dimensions are assumed to be independent. Sharpening each factor to load a few variables is its effect. Kline (2015) describes this technique.

Source: Kline 2015

Varimax rotation works by rotating the factors, keeping them orthogonal, so as to maximise the variance of the squared loadings within each factor, which pushes loadings toward either high or low values and away from intermediate ones. This produces a clearer structure in which each factor is defined by a distinct subset of variables. So varimax rotation works by seeking a configuration of uncorrelated factors that gives the sharpest contrast between high and low loadings, so that variables load strongly on one factor and weakly on others, which makes the factors easier to interpret while preserving their independence.

Source: Kaiser 1958

Varimax rotation is used to improve the interpretability of a factor solution when the factors can be treated as uncorrelated, since it produces a simple structure with each variable loading mainly on one factor, making the factors clearer to interpret. So varimax rotation is used as a default orthogonal rotation to clarify factor solutions, valued for yielding distinct, independent factors that are easy to interpret, which is why it is common when the constructs are assumed unrelated or independent factors are wanted, though when the underlying factors are expected to be correlated, an oblique rotation may give a more realistic solution.

Source: Kaiser 1958

A varimax solution is characterised by uncorrelated factors, since the rotation is orthogonal, and by a simple structure in which each variable tends to load strongly on one factor and weakly on the others, with loadings pushed toward high or low values. So a varimax solution has independent factors and a clear pattern of loadings, which makes it straightforward to interpret, with each factor defined by a distinct set of variables, and this simplicity and independence are why varimax is widely used, though the assumption of uncorrelated factors is a limitation when the underlying constructs are in fact related.

Source: Kaiser 1958

Varimax rotation is appropriate when the factors can reasonably be treated as uncorrelated, or when independent factors are desired for interpretation or for producing uncorrelated factor scores. So varimax rotation is appropriate where the assumption of independent factors is acceptable, giving a simple and interpretable solution, which is why it is a common default for orthogonal rotation, though it is less suitable when the underlying constructs are expected to be correlated, in which case an oblique rotation that allows the factors to correlate would provide a more faithful representation, so the choice depends on whether independence is a reasonable assumption.

Source: Kaiser 1958