I think in fairness you're merging two areas of studies.
Mr. Neufeld quotes from his conformity audit. I'll take you through that. It's in annex C of his report. It was an audit ordered after the Etobicoke Centre byelection Supreme Court case. That conformity audit was conducted with 1,000 random polls from the 2011 election, 10 from Etobicoke, and then 50 from the three byelections in 2012, including my riding of Durham in Ontario. So a total of 1,160 polls were randomly sampled.
We've heard a lot about vulnerable groups being impacted. That's not part of the conformity audit. The vouching that was suggested to occur at a 95% confidence level was 120,000-plus vouching transactions.
Mr. Neufeld reported that of those transactions, 95,000-plus had errors, serious ones right through to some cases with multiple errors giving a rate of 42% to 80% errors. Do you think that's an acceptable error rate for a G-7 country, Professor Thomas?