The performance of information retrieval systems varies significantly by test topics. Even for those systems that have performed well on average, the results for some difficult topics are still poor. Previous studies have revealed that different optimization techniques should be used for those difficult topics. However, a prerequisite of the…

A method is proposed for overcoming the problem of multicollinearity in multiple regression equations where multiplicative independent terms are entered. The method is not a ridge regression solution. (JKS)

Problems associated with the use of gain scores, analysis of covariance, multicollinearity, part and partial correlation, and the lack of rectilinearity in regression are discussed. Particular attention is paid to the misuse of statistical techniques. (JKS)

Problems in measuring change abound. For many purposes, a pretest score is best treated as any other measure obtained at the time of the pretest. Where appropriate, regression analyses that treat the pretest no differently from other independent variables, and the posttest as the dependent variable avoid many difficulties introduced by gain…

With even the simplest bivariate regression, least-squares solutions are inappropriate unless one assumes a priori that reciprocal effects are absent, or at least implausible. While this discussion is limited to bivariate regression, the issues apply equally to multivariate regression, including stepwise regression. (Author/CTM)

The degree to which the chosen units of analysis are likely to produce spurious findings in staged combinations of multiple linear regression procedures are examined. The effects of grouping variables (e.g., classroom, school, and school district) on Procedures such as Coleman's semipartial regression and Mayeske's commonalities, in light of…

The relative precision of four methods of estimating missing data in principal components analysis was investigated. Artificial data with known characteristics, obtained from Cattell's "Plasmode: 30-10-4-2," was used with one third of the data on half of the variables being systematically eliminated. The four methods of missing data estimation…

Addresses three issues concerning the assessment of the overrating and underrating of academic competence: (1) the impossibility of separating effects of overrating and underrating from effects of perceived and actual competence; (2) the questionable validity of Connell and Ilardi's method; and (3) the proposal of a new method and its implications…

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…

Many topics in aging research address questions about group differences in prediction. Such questions can be viewed in terms of interaction or moderator effects, and use of appropriate methods to test these hypotheses are necessary to arrive at accurate conclusions about age differences. This article discusses the conceptual, methodological, and…

The uses of data on siblings for educational and psychological research purposes are explained. Examples are based on empirical data. Five types of research problems, using sibling data, are explored: (1) statistical control of family background variance in psychological or educational studies which use analysis of covariance (ANCOVA) or multiple…