Some observations on multivariate techniques in landscape evaluation

Dearden P. and Rosenblood L. (1980) Some observations on multivariate techniques in landscape evaluation, Reg. Studies 14, 99–110. Statistical methods of landscape evaluation involving multiple-regression techniques are gaining in popularity. It has been suggested that the measurements of landscape elements be factor analysed before input as independent variables into the regression equation. This paper describes an experiment undertaken with the same data in the same area to ascertain whether different visual-quality maps are produced using just a multiple regression, or one preceded by factor analysis, and if so, whether the use of factor-analysed independent variables is to be recommended. It is concluded that the visual-quality map produced using individual landscape elements is the more valid.