An Analysis of the Charter Vote

March 24, 2018

The March 13 municipal election asked 2 questions of the voters regarding the charter. The ballot actually had 4 questions, but the first two were thrown out by a judge for being misleading, and were not counted.

Question 3 asked if term limited council members should be allowed to run again after sitting out 3 years, and question 4 would replace our “majority wins” system with a plurality – “highest vote total wins” system.

The results by precinct are shown on the maps below – click on the precinct for the individual results.

For a March election without candidates, the turnout was substantial, with about 4700 casting their votes. There were a couple of interesing anomolies about this election though – absentee ballot totals were much higher than normal, and at the polls, the number of “Under-Votes” was an excessive 3.4%.

We have a theory about the undervotes. (It is somewhat technical, so bear with me.)

An “under-vote” is when someone votes a ballot with nothing filled in. At the precinct, the optical scanners are supposed to only flag this as an error if NOTHING on the ballot is filled in – skipping a single question will be reported as an under-vote by the SOE in the results, but the scanner doesn’t flag it and the poll workers wouldn’t know about it.

In this case though, both questions had the same number of under-votes, implying that the ballots were blank and should have been flagged. When the scanner detects this condition, it beeps and prints on the tape “UNDER-VOTE” and spits the ballot back out. The inspector working the scanner is then expected to ask the voter if they intended to submit a blank ballot. If it was a mistake, the voter takes their ballot back and marks it. If the voter intended to submit it blank, then the inspector pokes a key in a hole on the back of the scanner to “OVERRIDE” the error and accept the ballot. We asked an inspector at one of the precincts with a number of under-votes if there were any overrides – and they couldn’t recall any.

Our theory is this:

Because of the short time to react after the lawsuit, the Supervisor of Elections did not have time to program the scanners to ignore the front page of the ballot (questions 1 and 2). Therefore, since 1&2 were on page 1 and 3&4 were on page 2, it is likely many people, knowing there were only 2 questions to answer, filled in the front page (1&2) and did not notice there was a back page. The scanners should have caught this but apparently accepted 1&2 as constituting a “non-blank” ballot, and did not flag it. As a result, about 160 people (enough to change the results on question 4), were possibly disenfranchised and not afforded the option to fix the error.

This is a “hanging chad” situation regarding the “intention of the voter”, particularly since it may have affected the outcome. There is a paper trail if this is the case – the tapes for each of the optical scanners are supposed to be saved (although I don’t know for how long). The tapes would show whether any “undervotes” were flagged and if that count equaled the 161 reported under-votes in Palm Beach Gardens.

The absentee ballots numbers were also interesting as 46% of the voters cast their ballots that way. This is significant because the absentee “YES” margin was 9 points higher on question 3 and 7 points higher on question 4. Without the absentee votes, question 4 would have lost handily.

Why do I mention this? Because all the questions were misleading and until the lawsuit and the little bit of campaigning that was done, most people were naturally drawn to the the “YES” answer. The absentee ballots of course went out before either of those things occurred.

The following table shows how the absentee ballot counts and under-votes fell during the last few elections.  Note that the high under-vote count in 2016 was because the municipal candidates appeared on the same ballot as the presidential primary candidates and many people only vote for President, ignoring the rest.

 

Year Votes Cast At polls Absentee % absentee Undervotes %Undervotes
2018 4731 2569 2164 45.7% 167 3.5%
2017 5240 3486 1754 33.5% 0 0.0%
2016 15257 12465 2802 18.4% 1958 12.8%
2014 7167 5556 1611 22.5% 79 1.1%
2013 4217 3390 827 19.6% 1 0.0%

Question 3 – Run Again?

Question 3
Strong Yes Weak Yes Very Close Weak No Strong No
Click the precinct on the map for vote totals.

Precinct Registered Cast Turnout YES NO YES %
1186 1457 163 11 78 79 50
1188 897 63 7 34 25 58
1189 96 11 11 9 2 82
1190 3391 245 7 159 83 66
1192 1454 124 9 78 40 66
1194 2029 401 20 315 74 81
1228 446 65 15 38 26 59
1238 1680 301 18 148 146 50
1240 2506 301 12 197 97 67
1242 2559 400 16 266 113 70
1244 1567 206 13 125 76 62
1246 2397 275 11 115 153 43
1247 228 32 14 27 5 84
1248 1495 204 14 103 95 52
1250 62 7 11 1 5 17
1252 2320 500 22 272 197 58
1254 455 32 7 15 17 47
1260 1293 172 13 70 98 42
1266 463 3 1 1 2 33
1268 295 8 3 6 2 75
1272 1911 189 10 88 94 48
1274 1585 145 9 49 88 36
1280 399 30 8 19 10 66
1284 2219 153 7 117 30 80
1288 40 5 13 3 1 75
1290 2130 224 11 127 88 59
1292 71 17 24 13 4 76
1296 713 34 5 24 9 73
1324 1282 73 6 42 29 59
1326 553 68 12 32 31 51
1340 12 3 25 3 0 100
1352 2210 161 7 91 68 57
1360 891 114 13 71 40 64
1372 153 2 1 1 0 100
TOTAL 41280 4731 11 2737 1827 60

Question 4 – Plurality or Majority?

Question 4
Strong Yes Weak Yes Very Close Weak No Strong No
Click the precinct on the map for vote totals.

Precinct Registered Cast Turnout YES NO Yes %
1186 1457 163 11 75 82 48
1188 897 63 7 25 34 42
1189 96 11 11 9 2 82
1190 3391 245 7 134 107 56
1192 1454 124 9 71 45 61
1194 2029 401 20 260 128 67
1228 446 65 15 33 31 52
1238 1680 301 18 115 178 39
1240 2506 301 12 163 131 55
1242 2559 400 16 215 161 57
1244 1567 206 13 111 89 56
1246 2397 275 11 87 179 33
1247 228 32 14 18 14 56
1248 1495 204 14 75 122 38
1250 62 7 11 2 4 33
1252 2320 500 22 234 230 50
1254 455 32 7 16 16 50
1260 1293 172 13 58 110 35
1266 463 3 1 0 3 0
1268 295 8 3 5 3 63
1272 1911 189 10 79 104 43
1274 1585 145 9 51 86 37
1280 399 30 8 18 11 62
1284 2219 153 7 106 42 72
1288 40 5 13 0 4 0
1290 2130 224 11 109 106 51
1292 71 17 24 11 5 69
1296 713 34 5 23 10 70
1324 1282 73 6 41 30 58
1326 553 68 12 23 40 37
1340 12 3 25 3 0 100
1352 2210 161 7 71 87 45
1360 891 114 13 55 56 50
1372 153 2 1 1 0 100
TOTAL 41280 4731 11 2297 2250 51