Sampling of Constituency shows 90% error rate

Vote loss scandal as 90% names are ‘wrong’ (requires registration with

OK, I am taking this headline with a small pinch of salt, not having had any visibility of the methodology used, sample size taken etc. There are lies, damn lies and statistics after all. 52.5% of cats surveyed said that they knew that ;-). Also – what is the definition of “wrong” in this case?

However – if it is representative of the trend in the Register nationally then a whopping great problem exists that needs some sort of concerted plan to be put in place to address it.

My question is this… what study has been done to identify the potential root causes of this error?

One anecdotal cause I have heard about just today concerns people who are changing their names for reasons of marriage, divorce or other. The only form that is issued is a change of address form (or so I’ve been told). But if you have been co-habiting and have then married, you will not necessarily be changing address. Therefore if you opt to register in your married name (Ms. Jones becomes Mrs Smith at the same address) the process throws, for want of a better term, at total wobbly and it is possible that Ms. Jones would remain on the register while a new entry for Mrs. Smith is created.

This process probably worked fine back in pre-Celtic Tiger Ireland where people co-habited less than they do now before marriage. However it has failed to keep pace with the expectations of the Information Consumers and other actors in the process.

To the political classes… please stop playing political ping pong with this issue. Take action NOW to identify the root causes of these errors, address deficienies in process.

One good starting point would be to attend the Master Class on Information Quality that the Irish Computer Society and the IAIDQ are hosting in the RDS on Friday…

Short term actions that could be taken:

1) Invest in software (ideally a standard software tool across all Local Authorities) to clash Electoral register data with the Register of Deaths and the Register of Marriages to cut out noise and reduce the incidence of dead men voting or blushing brides voting early and often. This investment in a standard toolset could leverage economies of scale and would also standardise approaches across local authorities while supporting reuse of clean-up processes. Examples of such tools would include Informatica Data Quality (formerly Similarity Systems Athanor) -which is a IRISH made product (nice Celtic Tiger success story there).

2) Commence a root cause analysis review across ALL Local Authorities under the management of a central governance board to identify the process failures

3) Fix process gaps, introduce standardisation, amend processes, work practices, information management approaches etc. to enable the process to deliver a product that meets or exceeds expectations.

To build the business case for this work, I would call on the statistically minded politicos to work out what the cost to the State is from process failures in other areas that rely on the Electoral Register – for example the Jury Duty process – how many letters are sent to people who are dead or no longer exist at an address each year.

From the data complied by Ogra Fianna Fail in Trinity up to 90% of the spend in that downstream process could be wasted.

What is the COST to the Exchequer of doing nothing to improve the quality of these processes on a continuous basis?