This examination of data selection preceding multivariate analysis compares results grained with "gentle" and "draconian" variable elimination. To acquire comparable results, two stages of statistical exploration into an integrated model of motivation, learning strategies, and quality of teaching were used. The goal of the paper was to select variables that were as powerful as possible and to use them to produce a new factor model, applying both standard linear (Statistical Package for the Social Sciences [SPSS]) and Bayesian (BAYDA) statistical software. The second stage of the research, planned for the future, will involve efforts to produce sensible path models related to the theory. The study of motivation, learning strategies, and teaching used research in a Finnish telecommunications company in which 156 respondents answered a questionnaire with 135 items. The statistical analyses used for variable selection are described. Research results indicate that both teaching style and subject material can influence study motivation and the use of learning strategies. Learning material can be described with respect to the degree that broad theoretical perspectives are developed, and the personal meaningfulness of the subject material or the drawing of logical conclusions from detailed evidence is emphasized. The new statistical methods described, especially the nonlinear approaches based on Bayesian networks, enable the application of statistical analysis to the study of student motivation and self-regulated learning. The SPSS syntax is attached. (Contains 19 tables, 6 figures, and 87 references.) (SLD)