The Use of Influence Diagrams in the Design of Spreadsheet Models: an experimental study

Abstract

The spreadsheet is one of the most commonly used personal computer applications. Many studies have shown that a very high percentage of spreadsheet models contain serious errors. The use of influence diagrams, which are a visual representation of the relationships among the variables in a problem, to plan the development of a spreadsheet model may reduce the likelihood of the spreadsheet model containing an error. This paper describes an experimental study that tests whether the use of influence diagrams improves spreadsheet modelling practice. The results of the study suggests that influence diagrams do help users to better understand problems, and as a result reduce the likelihood of omitting important variables from spreadsheet models.

Beyond Regulatory Compliance for Spreadsheet Controls: A Tutorial to Assist Practitioners and a Call for Research

“…In one empirical lab study, participants were trained on a specific design approach for a specific error type (using Data Flow Diagrams for linked errors), which resulted in lower error rates for that error type [Janvrin and Morrison, 2000]. Another study showed that the group creating influence diagrams had fewer omission errors than the control group [O'Donnell, 2001]. However, in addition to requiring advanced training, these formal specification and development techniques may impose such a high cognitive load on the enduser, who is not used to developing models in a structured manner, that gains in the error reduction typically associated with the approach are offset by new errors created by cognitive processing difficulties.…”

Introducing a Taxonomy for Classifying Qualitative Spreadsheet Errors

“…During the operational usage phase of a spreadsheet, quantitative errors can result from accidental deletion, alteration of formulas or from incorrect inputs provided by end users. While quantitative errors can occur at any phase of the spreadsheet life cycle, research lab studies have focused primarily on evaluating techniques to minimize or detect quantitative errors during the development phase (Galletta, Hartzel, Johnson, Joseph & Rustagi, 1996;O'Donnell, 2001;Panko & Halverson, 1997).…”