Statistics Can Help Ensure Accurate Elections

The American Statistical Association’s Science and Public Affairs Advisory Committee has recommended that post election audits have at least a 90% level of confidence.

Election officials need to make sure the person elected winner is the person the most voters want..Election results are most trustworthy when the entire election process can be audited, not just the vote counts…the audit should have the statistical power to trigger additional action at least nine out of 10 times when the wrong winner is declared.(emboldening added)

What is significant with this committee is that it has recommended a specific level of confidence where other reports have used 90% or 95% as examples. For instance here and here.

As I have emphasized here and here Connecticut law is, in my opinion, inadequate because it over audits some races and it is woefully inadequate in other races. For example, in races such as state representative, senator, or small to mid size municipal races, the probability of detection would be 2%-4%. or one in 50, or one in 25, which is a far cry from nine out of 10. (Note: CT Law specifies auditing 10% of races selected randomly, which is quite different than the confidence %, a measure which represents the probability an audit confirms the correct candidate has won an election)

The committee goes on to cover other areas where statistics can be used to increase the integrity and confidence in the whole process. Read the entire report here.

The President of the association, however, recommended an even higher standard of 99% confidence in a letter to Senator Feinstein who, along with Senator Dodd has proposed woefully inadequate legislation at the federal level.


Leave a Reply

You must be logged in to post a comment.