THREE WAY PRESIDENTIAL RACE A DEAD HEAT, RUNOFF POSSIBLE

---

FLASH POLL FOR DECEMBER 2024 PRESIDENTIAL ELECTION
Results and Analysis
Written by PhDre

---

---

The EBC commissioned a poll on 12/18 to mark the entrance of the HEM / Sopo ticket. This marks the first poll conducted with the current three candidates in the race, and 30 Europeians responded to the poll.

QUESTION 1 - HOW WOULD YOU CHARACTERIZE YOUR SUPPORT AT THIS TIME

30% of respondents are undecided, with Grea Kriopia beating out the latecomer HEM 30% to 27%. Vor comes in third place with 13%.

QUESTION 2 - (PUSH) HOW WOULD YOU CHARACTERIZE YOUR SUPPORT AT THIS TIME

Removing the "Undecided" option forced respondents to indicate support for one of the tickets. The three way race seems incredibly very likely to head to a runoff according to this push question. Note that of the Undecided voters from Q1, 44% went for Vor, 33% broke for HEM, and 22% supported GK. Clearly, undecided voters will play a key role in the shape of the race over the next 48 hours.

QUESTION 3 - SUPPORT LEVEL

Respondents were asked to indicate how strongly they supported their candidate, with 1 = UNDECIDED and 5 = STRONGLY SUPPORT.
The average "supporter" of each candidate (determined by answer to Q2):

• Grea Kriopia: 4.08
• Vor: 3.57
• HEM: 3.36

RUNOFF SCENARIOS

Presented below are the various possible runoff scenarios should no candidate reach a majority on tomorrow's Presidential Election ballot.

Editor's Note: An earlier version of this story stated that 3% of undecided voters in Q1 supported HEM in Q2. This was a typo, the actual result is 33%. The writer responsible has been fired.

 

[Gem]

This is fascinating, and definitely not what I expected! Thank you for doing this poll, Dre!

 

[John Laurens]

THE ELECTION OF 1800! CAN WE GET BACK TO POLITICS?

 

[HEM]

I mean, it's kinda fascinating. There is something for every ticket to like in here.

Obviously as the late entrants, being where we are is pretty astounding. But our support is (understandably) the softest by far.

If you're Vor, you're a little worried about the initial head-to-head, but if you can get to the runoff you are in a stronger position than earlier polling indicated.

If you're GK, you have the most staunch supporters and are in the most enviable electoral position over all.

 

[PhDre]

I mean, it's kinda fascinating. There is something for every ticket to like in here.

I didn't want to editorialize but yes that's exactly my take away, I wouldn't hate to be any ticket ATM and each has a very viable path to the Presidency. Now lets hope election day goes smoothly and Russia doesn't try any funny business ..

 

[sanjurika]

bug if true

 

[Prim]

So, I'm curious... how can we get better and more representative polling on elections? Seems like this was a big miss here, pretty wide swing. I understand that small communities are tougher to get truly representative polling, it's basically more of a mini-vote in poll form, but we consistently get 30 or less responses to polls, but 50-60 votes in elections...

 

[John Laurens]

So, I'm curious... how can we get better and more representative polling on elections? Seems like this was a big miss here, pretty wide swing. I understand that small communities are tougher to get truly representative polling, it's basically more of a mini-vote in poll form, but we consistently get 30 or less responses to polls, but 50-60 votes in elections...

Calculate how strongly low propensity voters break for incumbents and then apply a shift based on that.

This isn’t the first polling miss that’s happened, the last Senate General Election there was a similar miss

 

[Gem]

So, I'm curious... how can we get better and more representative polling on elections? Seems like this was a big miss here, pretty wide swing. I understand that small communities are tougher to get truly representative polling, it's basically more of a mini-vote in poll form, but we consistently get 30 or less responses to polls, but 50-60 votes in elections...

I think one factor in how ultimately lopsided this election was had to do with the lopsided GOTV effort. Granted, I don't think that made up for the entirety of the polling error here, but GK and Sin definitely outstripped their opponents by quite a bit on outreach, and the people likely to respond to these polls would probably be the least easily swayed by that.

If I had to guess, I'd say an outsized proportion of the 38 people who didn't take this poll went to GK for that reason.

 

[Comfed]

It's a truism that Presidential polls always skew towards only representing the 20 to 30 active citizens who follow the Presidential race and develop an informed opinion. The remaining 30 to 40 voters do not follow regional goings-on actively enough that they will fill out an opinion poll. That issue was probably compounded in this case by the short time both opinion polls were open -- the ENN poll wasn't even open for 24 hours.

I agree somewhat with JL, but I think it was more complex than that. In my experience, the citizens of a region who are the least steeped in its domestic politics are, in general, the more likely to support mainstream positions; for example, in supporting incumbent or conventional (not to use that term in a prejorative sense) candidates for office. I would say that descriptor certainly applies to GK, who has been (deservedly) a highly popular Vice President, and cabinet minister before that.

But I don't think the result can be boiled down to just an incumbency effect. "Low propensity voters" (i.e. those not highly steeped in domestic politics) are also more responsive to candidates who reach out to them personally, which is why GOTV is so effective. I would also say that GK is probably the most recognizable candiate among low-engagement/information voters; both by virtue of being the sitting Vice President, but also because of things like personally welcoming every new citizen to the Discord server during her Vice Presidency. I certainly don't mean to suggest that she did that as some sort of political ploy -- it's just a good thing to do -- but I imagine it helped in the election.

So I don't think it can be boiled down to a simple "incumbency advantage" number. Rather, I would say that polls can only be more representative if they are actually reaching more people. A possible way to do this would be to send the poll to every citizen via forum PM, Discord message, or telegram. Of course, that's a lot more work than creating a Google form and posting the link in the Grand Hall.

 

[John Laurens]

I’ll go back and pull some data but I did predict the shift correctly based on previous trends but never put data behind it, so let me check

 

[PhDre]

So, I'm curious... how can we get better and more representative polling on elections? Seems like this was a big miss here, pretty wide swing. I understand that small communities are tougher to get truly representative polling, it's basically more of a mini-vote in poll form, but we consistently get 30 or less responses to polls, but 50-60 votes in elections...

I don't think this is a huge miss. Polls understated GK moderately and overstated HEM support moderately.

You could weigh voter characteristics based on demographic information from the Euro Census.

You could apply a "low propensity shift" which would be similar to weighting for voter characteristics.

I also wonder how much it matters that we have some polling miss. Like in US elections polling misses were used to propel distrust in the media or generate conspiracy theories about the quality of election security. Do we have a fundamental problem with being off? Do we want to be more explicit that there's an unreached electorate? Do we want to ping every single voter in the region for a poll two days before an election?

Finally it seems like the GK campaign ground game was significantly deeper than either opponent, if true it makes sense that it would make the difference in a relatively tight race.

 

[John Laurens]

I also wonder how much it matters that we have some polling miss

I think deep down we all just wanna be right

 

[GraVandius]

It’s unquestionably GOTV efforts in my mind. GK sent a message drawing eyeballs to their platform a few days ahead of the election and then followed up with a prompt message as polls opened. The only other GOTV message I received was a quasi 11th hour one from Rand. I’d bet money HEM’s ticket did none. It takes very little effort to swing people on the fence and there’s no better way to get support than clearly showing you want the job.

 

[Prim]

So, I'm in agreement about the GOTV commentary. I am really wondering though, how we can improve our polling to try to encompass some of these factors... low engagement voters (who are equal in size to our high engagement voters)... and GOTV efforts by campaigns?

I don't like leaving things at "well, there are holes in our polling that we just can't patch."

Should we theorize about likely voter models and what that could look like? Incumbency factors? Ask campaigns about GOTV efforts akin to analyzing ad spend in IRL campaigns?

 

[Lethen]

Do we think the other half of the electorate who don't take polls read the platform threads? If so could it be as simple as where we post polling threads? I wonder if we'd see a change in response rate if we also posted poll threads in the Oval Room

 

[Gem]

So, I'm in agreement about the GOTV commentary. I am really wondering though, how we can improve our polling to try to encompass some of these factors... low engagement voters (who are equal in size to our high engagement voters)... and GOTV efforts by campaigns?

I don't like leaving things at "well, there are holes in our polling that we just can't patch."

Should we theorize about likely voter models and what that could look like? Incumbency factors? Ask campaigns about GOTV efforts akin to analyzing ad spend in IRL campaigns?

At the end of the day, I think the only way we could truly nullify this factor would be to reach out to people the same way GOTV campaigns do—by going down the citizen registry person by person and sending everyone a DM or telegram asking them to respond to the poll.

Though that approach could (rightfully) raise some questions about whether or not it's worth it to go through the effort and potentially overload people with messages around election season just to get more accurate polling data.

 

[Monkey]

Could we ask a question asking who they’ve received a GOTV from and weight the responses according to that?

 

[Gem]

Could we ask a question asking who they’ve received a GOTV from and weight the responses according to that?

Ooh, that's a really interesting idea.

We would probably have to have people rank the gotv they've received from each candidate, because it's not a true or false—not all gotv campaigns are created equal. Some candidates send more and better gotv messages than others, even if the others do send one or two uncompelling ones.

And we'd probably have to do this for a few elections before the predictions we make based off of it start to be accurate. It would be hard to quantify and account for specifically how much gotv matters without a fair bit of data to use as reference.

But at least me personally, I'd be excited to try that out!

 

[Kazaman]

The poll itself seemed fine. The problem was the article's (and headline's) inference, which was a really risky leap from what the poll said. More conservatively, it could've just inferred that HEM's support mostly came from Vor's camp, which seemed to be about right.