1. Dollar , Charles M. and Jensen , Richard J. 1971 .Historian's Guide to Statistics89 – 90 . New York
2. In algebraic terms, the derived unfolding solution is anadditivelinear model, where the regression-covariance solution is amultiplicativelinear model. For the Euclidean unfolding case, the total vote for candidate i over a set of wards is the sum of the candidate-to-ward distances on all the wards: where the xiare the coordinates for the candidate points the yiare the coordinates for the ward points and the k are the dimensions used in the solution. For the regression vector case, the same vote is a weighted sum: where aijis the loading for candidate i on dimension k, and bjkis the weight for ward j on dimension k. The main consequence for visual representations is that the aiand the bjfor the regression case must be scaled in different coordinate systems, while the Xiand yjfor the unfolding case are scaled in a single system and can be described in the same units of measure