SPILO: Issues with RTM January Vote

Submitted by Michael Spilo, views are his own

Before Covid, the RTM voted with paper ballots.  RTM members would mark their vote in ink by their name and then pass the ballot to the next member of their district.  The District Leader (DL) would review the card and ask anyone who didn’t vote if they wanted to vote.  Sometimes the DL would ask if they voted as intended, say if a “no” vote meant you were in favor of the item.  If a voter made a mistake or changed their mind before the DL handed in the ballot, they could change their vote.  If there was a dispute after the meeting the ballots served as an audit trail.  Before Covid there was no time limit on voting.

In December the RTM implemented an electronic voting system (EVS) in part to streamline voting by remote users. In hindsight, the implementation of the EVS had several unintended drawbacks:

1. There are periods during which a vote may NOT be counted by the system because the system operator was not ready

2. There was no way to confirm WHAT vote was cast, and it was nearly impossible to confirm IF a vote was cast because the screens were illegible even up close

3. It is very easy for voters to change their vote multiple times – an accidental key press on the remote changes the vote, as do errors in the wireless transmission

4. Any issues like a lost connection, a dead or broken remote, etc. Have to be resolved within the 1 minute allotted for voting, even though voters had no way of knowing these issues occurred

5. The implementation did not provide any independent audit trail

For example, issue number 1 occurs when the RTM Moderator announces that the vote is open, but the system admin was not yet ready.  RTM members start voting but their votes cast don’t count. Issue number 2 occurs if you cast your vote and then accidentally “pocket” clicked another key on your clicker. 

Looking back at the votes in both December and January it seems likely that at least issue 1 described above happened on every vote.  Only when the vote tally for CTCL funding showed 104-101, did RTM members check, and it became apparent that several votes had not been counted and some were not cast as intended. Both Yes and No votes were affected. RTM members objected immediately.

The question now is what to do about that vote.  The answer is clear. Correct the voting process by allowing voters time to confirm and correct their FINAL vote and re-do the vote on CTCL funding. 

There have been a lot of immature responses being voiced online.

Some are crying “election denial” and “no do-overs” but the RTM has regular “do-overs,” and often re-considers items that were previously voted.  Examples abound: Bumpouts on Greenwich Ave, Police on Greenwich Ave, Northwest Fire Station (countless votes), Sound Beach Ave changes (3 votes). 

And while this is not an election, (the RTM has monthly votes) even in an election, when a voting system fails to count the votes, a re-vote is sometimes the only way to address the situation. [1]

In Connecticut voter intent is part of any “recount” or “recanvass” of a vote. Several RTM voters were disenfranchised yet no effort was made to identify the voters’ intent. [2 p 12] And as there is no independent audit trail, a recount is not possible.

Some claim this is a “slippery slope” and will lead to recounting every close vote. But again, the RTM regularly re-votes on issues and there are many long standing procedural methods to appeal or review any vote. In this case, the RTM Moderator indicated via e-mail that a motion to “rescind” was her preferred form of an appeal, so that’s what the LCC did.

Second, no one in the past has ever filed an affidavit that their vote was miscast, in this case several did. 

Third, the paper ballots served as an audit trail, and these were checked and occasionally a problem was found and corrected but never anything like the 11 potential missing votes as on the CTCL issue.

Lastly, the problem was introduced with the EVS and once the EVS process is fixed, other votes will presumably not have such large discrepancies.

Some say, “if we re-vote this, we must re-vote all.” Not so. There was only a protest over one vote and only affidavits signed for one vote, and only one vote had enough missing votes to affect the outcome.  So only one vote needs a recount.

Others say, “let the BET deal with the spending.” But spending isn’t the only issue.  Many feel that the Registrars are in over their head when it comes to technology and will be led astray by CTCL and their close connections to Microsoft, Facebook and other tech companies.  The RTM EVS rollout is an example of why such decisions should undergo scrutiny – a similar mishap in a general election would be much more costly and divisive.

Some say, “user error – too bad for them.” But to err is human and a good system is designed to minimize user error by providing feedback to the users and an opportunity to correct any errors. 

The silly one-liner objections are endless.  The bottom line is that the new voting system did not operate as intended.  There was only one close vote whose outcome may be affected by voting system discrepancies. 

We can fix the problem by adjusting the voting process to let RTM Members confirm their FINAL vote and then taking one additional vote in March. The re-vote could go either way.  Hopefully we won’t see political maneuvering and multiple votes attempting to derail what is undoubtedly the right course of action.

By Michael Spilo,
The views presented are my own.

1. https://www.politico.com/magazine/story/2018/12/18/history-election-do-overs-223313/

2. https://portal.ct.gov/-/media/SOTS/ElectionServices/Handbooks/Recanvass-Manual-2012.doc