Authors
N. Davies & Chris C. Ikin
Abstract
While there are rumors of multimillion-dollar mistakes being made through the incorrect use of spreadsheets, it has been difficult to confirm these reports. In order to help determine the accuracy of spreadsheet use and the quality of the controls applied to them, the Tasmanian State Institute of Technology did a study of firms and some of their spreadsheet applications.
The objective of the analysis was to isolate 2 types of faults - existing errors and potential errors. Interviews with company personnel indicated that:
- Most believed their worksheets did not contain errors.
- Manual audits are rarely performed.
- The most common form of training is "on the job."
Numerous problems were found in the analysis of most worksheets. Although not as strong a control as error prevention, error detection routines can limit faulty spreadsheet output.
Sample
There seem to be two solutions to the problems identified in this study: preventing errors and locating those errors which do occur.
Error prevention:
- Design worksheets in logically related blocks.
- Use appropriate documentation.
- Use range names instead of cell addresses.
- Numbers should never appear in formulae unless the number is a constant (number of days per week, etc).
- All cells should be protected except input cells.
- Use input forms which facilitate accurate and complete data and act as a valuable audit trail.
Error detection:
- Conduct reasonableness reviews.
- Include controls to check that data is entered correctly and formulae are operating as intended.
- Use spreadsheet audit software.
- Involve the internal audit group.
Publication
1987, In the black, Volume 57, Number 11, December, pages 54-56
Full article
Not available
