Halloween, Elections and Spreadsheets

Halloween, Elections and Spreadsheets


Halloween beckons shortly, and perhaps it’s time for some spreadsheet horror stories! And in this US election year, especially ones with an election flavour. The EuSpRIG horror stories web page can be found at:


and contains two reports on voting miscounts, and three more in their archive page, and you can easily find more by searching. For example, because of a mis-entry on a spreadsheet, one miscounted vote reversed the Anglican Church of Canada’s rejection of same-sex marriage, July 12, 2016.

Sometimes it’s simply typos, or what are known as “fat fingers”, sometimes entering numbers in the wrong place, in other cases totalling formulas either double counted or omitted numbers. That’s why those common mistakes are specifically addressed in the Spreadsheet Safe Syllabus, and a core part of the learning and training there.

The task of reconciling the integrity of data from external sources as complete, consistent and correct is a central tenet of the Spreadsheet Safe programme. The programme considers five ways in which totals can be checked, of which at least two important ones apply to the reported incidents given above. These self-checking routines or techniques are the advice to check for and correct errors in totalling caused by row or column insertion or deletion; and to check for double counted sub-total errors.

At the same time Spreadsheets themselves can be used as a parallel checking system to validate the results of large scale business systems, for examples company wide ERPs. Again an important and useful point, because the act of reconciling any discrepancies ensures that the data are looked at carefully, whatever its source. The advice to create cross-total checks should also apply here. This is a crucial consideration for those working in high stakes environments, and should be part of the routine controls.

It’s easy to blame junior team members, but all human error research shows that any person operating under speed pressure or overly long hours is much more likely to make mistakes.

Research by one of the leading acacdemic researchers in the area, Prof. Ray Panko from the University of Hawaii, has confirmed that spreadsheet users are no different from any other knowledge worker in that respect.

Back to the hustings, we need then to recognise when the pressure is on to produce the election results overnight after the polls close, errors may happen. Recognising this means adequate time and resources are always necessary.

Patrick O’Beirne
26 Oct 2016

Spreadsheet Safe Programme

Syllabus Reference: Reconcile the integrity of data from external sources as complete, consistent and correct.

No Comments

Post A Comment