This paper reports on attempts by the author to construct a theoretical framework of adult education participation using a theory development process and the corresponding multivariate statistical techniques. Two problems are identified: the lack of theoretical framework in studying problems, and the limiting of statistical analysis to univariate techniques. The process of theory development is divided into: (a) definition of concepts; (b) establishment of variable relationships; (c) specification of causal ordering of variables; (d) identification or determination of relationships among variables within the same causal system; and (e) estimation of change. The use of regression analysis and models in the final phase is discussed, as is also functional path analysis to determine whether the proposed set of interpretations is consistent throughout. The final results will be presented in a doctoral thesis. (author/pt)