An "undesigned" experiment is one in which the predictor variables are correlated, either due to a failure to complete a design or because the investigator was unable to select or control relevant experimental conditions. The traditional method of analyzing this class of experiment--multiple regression analysis based on a least squares criterion--gives rise to a number of interpretation problems when the effects of individual predictors are to be assessed. Some difficulties and their effects on the quality of information are discussed. Two methods are described in this report for improving the information obtained from the undesigned human factors experiment. One is to collect more information at a few data points selected at locations that improve the orthogonality of this non-orthogonal design. The other is to use a ridge regression analysis in place of the conventional least squares analysis, in which a slight bias is introduced into the data in such a way that the combined bias and variance error is smaller than the variance error of unbiased estimates from the least squares analysis. The ridge analysis produces more stable and meaningful regression coefficients. Computational aids--both references and complete computer programs--are supplied. (Author)