Authors
Simon Thorne & David Ball
Abstract
Most spreadsheet surveys both for reporting use and error focus on the practical application of the spreadsheet in a particular industry. Typically these studies will illustrate that a particular percentage of spreadsheets are used for optimisation and a further percentage are used for 'What if' analysis.
Much less common is examining the classes of function, as defined by the vendor, used by modellers to build their spreadsheet models. This alternative analysis allows further insight into the programming nature of spreadsheets and may assist researchers in targeting particular structures in spreadsheet software for further investigation.
Further, understanding the functional make-up of spreadsheets allows effective evaluation of novel approaches from a programming point of view. It allows greater insight into studies that report what spreadsheets are used for since it is explicit which functional structures are in use in spreadsheets.
We conclude that a deeper understanding of the use of operators and the operator's relationship to error would provide fresh insight into the spreadsheet error problem. Considering functional spreadsheet operator usage suggests the value of Example Driven Modelling for Decision Support Systems.
Sample
The first study into spreadsheet error was conducted by Brown and Gould for IBM in 1987. They found that 44% of the spreadsheets developed contained errors such as mistyping formulae.
Since this original paper, there have been many studies in spreadsheet error and the statistics reported from these studies varies from 30% to 100% of models with errors.
This table depicts some experimental studies with relevant error rates.
Publication
2006, EuSpRIG