Search

In an attempt to bridge the gap between academic understanding and practical business use, the Box-Jenkins technique of time series analysis for forecasting future events is presented with a minimum of mathematical notation. The method is presented in three stages: a discussion of traditional forecasting techniques, focusing on traditional techniques that relate to the Box-Jenkins model; a presentation of the key characteristics of the Box-Jenkins model; and a detailed demonstration of the method's steps. A discussion of the technical requirements needed to use the model, its cost versus that of other forecasting methods, and the potential areas of its application conclude the paper. The mathematical rationale is appended. (Author/DW)

Models of five alternative ways to approach the evaluation of organizational development (OD) projects are sketched: (1) ask the expert; (2) what's the target; (3) did we hit the target; (4) mid-course correction; and (5) continuous monitoring. Each model provides a different perspective on the problem of evaluation and each contains certain advantages and disadvantages. The five approaches are next put into organizational context by a survey of some major obstacles to effective evaluation studies. Persistent obstacles are found in the client, in the OD consultant, and within the OD activitv itself. In the final section, two issues are addressed that must be faced in any effort to evaluate an OD program--clarity of the role of evaluation and the audience of the evaluation. (Author)

It is recognized that there is a major gap between the promises of large data bases and optimization and simulation models and their actual ability to solve real world problems. This document describes a Generalized Data Base Planning System (GPLAN), currently being developed at Purdue University, that is proposed as a system to bridge this gap. A number of systems contain some of the components of GPLAN, and a survey of each system is presented with details of its contribution to GPLAN design. Research on two systems at Purdue is directly leading to the development of GPLAN. One application is concerned with the development of a regional water pollution control planning system; the other is an interactive information system design package. The authors suggest that GPLAN is as much a major advance over a Generalized Data Management System (GDMS) as a GDMS is a major advance over application programs with formatted sequential files. A bibliography is included. Diagrams may reproduce poorly. (Author/DN)