Authors
Polly Brown & John Gould
Abstract
Nine experienced users of electronic spreadsheets each created three spreadsheets. Although participants were quite confident that their spreadsheets were accurate, 44% of the spreadsheets contained user-generated errors.
A main purpose of this experiment was to test whether experienced users of spreadsheets make errors and, if so, what the nature of those errors are. Errors were common and this experiment may underestimate the actual error rate.
Causes of errors included:
- Participants pointing to the wrong cell(s) or typing the wrong cell address.
- Copying a formula from another cell which produced the wrong cell references.
- Incorrectly including data in a summation.
- Mistyping of formula operators.
- Misapplication of cell formatting.
- Copying a formula rather than the cell value.
- Substracting one value from another, instead of the other way around.
- Mistyping data values.
- Rounding an intermediate result instead of using a cell format.
- Failing to unprotect a cell.
Sample
Participants each created three spreadsheets. Forty-four percent of the spreadsheets had at least one error. All of the participants made at least one error.
Participants were generally "quite confident" that their spreadsheets were accurate. This confidence was just as high on the spreadsheets that actually contained errors as it was on the spreadsheets that did not contain errors.
Publication
1986, Proceedings of the Human Factors Society, Volume 30, number 1, September, pages 29-32
Full article
Not available
