The Count

I’ve been quiet here lately. March is generally my busiest month academically, in April I had several commitments in the United States, and when I returned to England at the end of April, I had the end of the year grading crunch (which I finished up just yesterday) and another week long morning run on BBC Radio Devon. Most weekends in England are occupied with my daughter, who craves constant interaction (while I am utterly unproductive during those weekends, they are my favorite times ever). When work backs up, sadly, my pattern has been that LGM is one of the first things to go. That, and beer.

A relatively new commitment is my activity with the Labour Party. I’m a member (they give you a card and everything) and I guess what you’d call an activist. During the run-in to the local and European elections held last week, I wanted to get involved in and familiarise myself with as many aspects of the operation as possible. This included stuffing envelopes, phone canvassing, leafletting, knocking on doors talking to people, getting out (our) vote on election day, and meeting a lot of very cool, like minded people. In the coming year, there will be much more of all of this, plus data analysis. I am impressed with the sophistication of the operation, but there are areas we can improve upon, and locally I’ll be making an evidence-based contribution in this area.

Also, I had the opportunity to serve as a count agent for Labour on election night and the next day during the formal count. Jonny Morris (pictured), an incumbent city councillor running for re-election, put my name forward. While not knowing what the hell it was, it sure sounded cool. I researched what the Electoral Commission has to say about agents’ rights and responsibilities, spoke to Jonny as well as two other Labour councillors, and following a fun, exhausting day working cleverly targeted GOTV in two of our target wards (from 05:00 to when polls closed at 22:00) went down to the Guild Hall in the center of town to do this thing.

The next 18 hours or so happened in two stages (separated by a few hours of glorious sleep). The 19 contested wards for Plymouth’s City Council were spread across three tables per ward in three different halls.  After the polls closed, the ballot boxes started to arrive. That night, the city’s job was to verify that the number of ballots coming out of a box equalled those that went in, and to stack them into groups of 25. My job was to gather data, both in terms of the overall picture in the ward, and ideally, by polling station. The ward I was responsible for had eight polling stations, and I was able to obtain a sample from six. This is the only time we can get this level of data as once the ballots are stacked, this information is lost for all eternity (the city keeps track of aggregate numbers by ward of course, and that’s the public record, but does not distinguish within the ward). This will be useful in the coming year.

There were three city counters per table, so I had nine in total to observe, by an large by myself. Our candidate was there for a couple hours but he was understandably exhausted considering all the effort he had been putting in, but at least I had three gregarious cool relaxed laid back UKIPers to keep me company. Having precisely zero experience with this, I had to devise a sampling method on the fly. The city was not obligated to show me the vote choice, just the ballot papers (as officially, my job was to observe for Labour that the count coming out of a ballot box was equal to the count that went in). Some counters were more cooperative than others; one entire table of three decided they would bundle the ballots face down, so I could not see a thing. Others would do so in such a way that I could see the bottom two candidates but not the top three, or the top two but not the bottom three; either were worthless in terms of obtaining a reliable sample. I did manage to find two or three who were transparent in their count such that I was able to obtain a decent enough sample. At one point, consulting with the candidate, I was able to say “this is where you are. these are good numbers. I trust these numbers”. I didn’t know it until that moment, but I always wanted to say that to a candidate.

There was one flaw in my sampling, and that’s the postal vote, which I think worked out to about 40% of the total turnout. Those were piled into their stacks of 25 first, and it proved very difficult to devise a reliable sampling technique for those.  I’ve retained everything of course and I’m looking at my postal numbers versus my ballot box numbers to see if there’s anything I can learn.

I was at the Guildhall from a little after the polls closed at 10pm until 1:30am. I was back at 12:30pm the next day for “The Count”, where the city takes those stacks of 25 and distributes them by party (into stacks of 25 that are paper clipped, then bundles of 100).  My job here, again, is twofold. I’m to ensure that no Labour votes end up in the piles of the opposition. In the ward I was responsible for, we had five candidates: Labour, the Conservatives, an independent candidate who runs every election, the “Trade Unionists and Socialists Against Cuts”, and, of course, UKIP. Meaning, five piles. Note, no Green candidate; and the Greens did quite well in a couple of the wards. Also, no Liberal Democrat, who stood candidates in merely four of the 19 wards. The only discrepancy I noted was twice one counter placed a Conservative ballot in the Labour pile. There was no Tory count agent present at the moment, only their candidate, who wasn’t observing this table. It’s not my job to do the Conservative’s job for them, so I said nothing. In each case, however, the counter caught his error before adding another ballot to the Labour pile, so in the end the Conservatives were correctly credited with those two votes, and the shaky foundations of the democracy held for another day.

The second job was to get an overall count of where we stand, and this is where the drama enters the narrative. We were well organised, with the candidate watching one table, me a second, and an additional Labour observer (who had done this many times before and was well on top of things) the third. We aggregated notes, and had confidence. My sample from the night before, 401 ballots out of ultimately nearly 4000 cast, indicated that it could be close, but we’d win and UKIP would finish second.

I was confident, that is, until I caught sight of the UKIP tally sheet, which credited them with over 200 more votes than we estimated. That our ward was called for a recount did not add to our confidence. I found solace in my numbers from the night before, and knew I had a much better handle on things than they had, but during the first of two recounts for the ward, I worked out that we had underestimated the UKIP vote. Telling the candidate that we didn’t have them where we thought was not my favorite part of the day / week / month. The situation was tense, and when we learned that turnout in the ward was 40%, we weren’t sure what to make of it. 40% is above average for local elections here, so did it mean that we got the boost higher turnouts usually afford Labour? Or did UKIP mobilize an atypical number of non-voters to their extreme right wing cause?

The second recount assuaged any concerns and the tension vanished. We had won. By 96 votes. The emotion was tangible. My estimates from the night before had the ward 34% Labour, 28% UKIP, 25% Conservative, 11% Independent, 1.5% TUSC. Again, those were without the postal ballots. The final result was 34.6% Labour, 32.2% UKIP, 23.4% Conservatives, 8.8% Independent, and 0.8% TUSC. My error was Labour -0.6%, UKIP -4.2%, Conservative +1.6%, Independent +2.2%, TUSC +0.7%.

Not too bad for one guy devising a sampling method on the fly, observing three tables, with nine city counters cooperating to varying degrees, while rejecting my estimates of the postal votes as being unreliable.

After that, we all gathered in the ballroom upstairs to watch the declarations of each ward (the source of the photo above), which culminated in re-elected Labour councillor Bill Stevens’ rousing speech that went viral on twitter here in the UK, and met with a decidedly impolite reaction from the UKIP contingent. Then, of course, it was off to the nearest bar, where every party, Labour, the Tories, Greens, TUSCs, even the odd Liberal Democrat were all in attendance, proudly wearing their rosettes and respective party bling.

Except UKIP. Even though they had elected three councillors to Plymouth City Council, they stayed away.