Election Recap: The Good, the Bad & What Should Be Changed About DFNYC's 2013 Endorsement Voting

Election Recap: Review of Democracy for NYC 2013 Endorsement Voting

A Report on Member Feedback: The Good, the Bad, & What Should Be Changed

In the past few weeks, we held our first round of voting for DFNYC endorsements in 2013 races. (Click here for the rules.) It was the first time  that we have run an endorsement vote with this many factors – several races, a candidate questionnaire with answers posted on our website, and only our second time doing online voting. 

This is our initial report on member feedback, along with the specific endorsement process, explained below.  More feedback, comments, or compliments? Email us at info-at-dfnyc.org.

1. The Ballot, Generally: Members reported generally that the ballot was clear and the task of voting was simple, straightforward, and not too time consuming.

2. Candidate Questionnaires: Members reported reading the candidates’ responses to our questionnaire, posted at our website, to help in their voting decisions.

3.  Technical issues:  For the mayoral race, there were two problems with the choices provided. (We’re not sure if this was a Survey Monkey glitch or our mistake in designing the ballot, but either way, we have solutions.)   Overvoting (voting for more than one candidate) was possible, and although "Other" was given as a possibility with a text box for write-ins, some members said they were not able to actually choose other. We became aware of this problem when it was mentioned in the comment section at the end of the ballot and we were able to fix the ballots to determine voter intent.  Our solution for the future is (1) we will ensure when setting up the ballot that “other” can actually be chosen, and (2) when we test the ballot, we will test it for both overvoting and write-ins.

3. Decision about Threshold for Endorsement:  The decision about the threshold percentage for endorsement should have been made and posted before voting began, and we should have been more proactive in getting member feedback on this decision.

4. Decision about Races:  We chose the races based on ones we had heard buzz about in the progressive community and also asked our members for feedback.  During that process, one more race was added to the ballot.

5. Decision about Timing:  Political groups that make endorsements grapple with the decision about when to have endorsement votes.  We chose to do so as soon as it was practical for us to hold a vote, as we wanted our endorsements to have impact, respect the time constraints of the candidates’ campaigns, perhaps help to influence DFA, and give us time to prepare before petitioning.  However, in the future, we should be more proactive in getting member feedback on the timing of endorsements. 

6. Endorsing Candidates that Did Not Respond to the Questionnaire.  In some races, the candidate who won the DFNYC endorsement did not answer our questionnaire, which raises the issue of whether DFNYC should be endorsing someone who does not answer our questionnaire, especially when other candidates put a lot of effort into their responses.  One the one hand, it seems fair to require questionnaire responses for an endorsement.  On the other hand, it may not be fair to our members to keep certain candidates off the ballot, when for some voting members, they have long admired the political activism of certain candidates, and the DFNYC questionnaire is just one of many factors influencing their vote.  If anyone has any idea for the future of how we might balance these two principles, please let us know: info-at-dfnyc.org.

~~~~~~~~~~~~~~~~~~~~~~~~~

The Election Process

IN elections before July 2012, we used paper ballots. We found this process to be difficult to organize and overly cumbersome for voters.  After getting member feedback, we decided to switch to online voting, and deal with the problem of anonymity by having DFAers in other states count the ballots.  Here is the step-by-step process for how it was done.

1. DFNYC leaders opened up an account on Survey Monkey, the "Survey Monkey Ballot Account" in April of 2013.  This was a completely separate account from our regular ongoing account where candidates were recently entering their questionnaire responses.

2. DFNYC Leaders designed the ballot. There was discussion among DFNYC leadership about what races to put on the ballot, and what candidates should be listed in each race, and the ability to have write-in candidates. Discussions were ongoing, but final decisions were made only after asking DFNYC members for their feedback in our newsletter as to races that should be included for endorsement voting.

3. Several DFNYC leaders "tested" the ballot by entering a vote that would not count. After this testing, we made a change: We made it more clear to voters when their vote was submitted. After the testing, we deleted all these test ballots.

4. When the ballot was finalized, DFNYC Leaders gave the username and password to the Survey Monkey Ballot Account to one of our out-of-state vote counters.  He promptly changed the password, so that DFNYC leaders would not have access to the account.

5. Voting opened on Tuesday, April 9th, and was announced in the email newsletter that morning. The email newsletter goes to all DFNYC members.

6. When voting opened, a decision had not yet been made as to the threshold percentage of votes needed for an endorsement. DFNYC leadership decided on a threshold of (a) more than 50% or (b) the most votes, and at least 10% points more than the closest opponent. This decision was posted before the voting period closed, but in retrospect, it should have been made and posted before voting opened, and with member feedback. (See above)

7. Voting was scheduled to close on Monday, April 9th, at midnight. In light of the events at the Boston Marathon and it being Tax Day, DFNYC leadership decided to extend voting one more day to Tuesday, April 16th at 5pm.  The reason DFNYC leaders chose 5pm and not midnight was to allow the vote counting to begin that evening, which was necessary in order to announce endorsements Thursday morning, while it would still be fresh in the political news cycle.

7. Determination of Eligibility & Anonymous Voting:

Our plan was for our vote counters to download all of the ballots (survey responses) to a spreadsheet and give us just the voter identity information (only those spreadsheet columns) so that we could determine eligibility.  Voters that had been to two events since 11/15/11 (fundraiser with Howard Dean) were eligible to vote. Then we would tell the vote counters what voters were not eligible, and they would delete those rows(Ballots and voter info) from the spreadsheet. Each of the two vote counters were to have the raw file and each were to make the deletions independently from each other. 

The process went according to this plan. The only difference was that DFNYC leadership requested and received, 24 hours before the end of voting, a preliminary list of people who had voted so that we could begin checking eligibility and do GOTV for active members that had not voted. On Tuesday, after the vote deadline, the vote counters gave us the full list of voters.  DFNYC leaders determined that just three voters were not eligible. The vote counters deleted these ballots (spreadsheet rows).

8. From the evening of Tuesday April 16th after voting closed through Wednesday, the vote counters separately calculated the number of votes for each candidate.  There were two ballots that were unclear as to the mayoral race, due to either a Survey Monkey glitch or a mistake in the way DFNYC leaders set up the multiple choice. During the vote counting process, the vote counters and DFNYC leadership worked together to reach out to the voters to determine voter intent.  (See above for our proposed solution for the future.)

9. The vote counters compared numbers and agreed on final totals on Wednesday afternoon, April 16th and informed DFNYC leadership. Members of DFNYC leadership called candidates on Wednesday night and results were posted on the website, DFNYC.org and published in a newsletter Thursday morning, April 17th.

10. On April 25, according to plan, the vote counters went into the Survey Monkey Ballot Account and deleted all of the ballots.  When that was completed, they gave control of the account (password) back to DFNYC leadership so that the account could be deactivated. A review of Survey Monkey policy indicates that it is not possible to retrieve deleted survey responses. (For our purposes, ballots are survey responses.)