500: Election

Explain xkcd: It's 'cause you're dumb.
(Redirected from 500)
Jump to: navigation, search
Someday I'll be rich enough to hire Nate Silver to help make all my life decisions. 'Should I sleep with her?' 'Well, I'm showing a 35% chance it will end badly.'
Title text: Someday I'll be rich enough to hire Nate Silver to help make all my life decisions. 'Should I sleep with her?' 'Well, I'm showing a 35% chance it will end badly.'


This comic was published the day after the 2008 presidential election in the US. Cueball has been closely following the quantitative aspect of the election for over a year and a half, and he seems to be relieved that it's over. Now that the election has passed, he does not have to follow the many different opinion polls, number-crunching analyses, and news clips about people like Joe the Plumber that he has kept close track of during the election season. As soon as he says this, however, he starts to search for information on the 2012 election, suggesting that his political obsession has not at all passed.

A list of the elements Cueball had been thinking about:

  • Opinion polls: These are simply surveys of voters' opinions on various issues and who they plan to vote for. They tend to be the primary source for predicting the outcome of elections as they can be created well in advance.
  • Exit polls: These are surveys conducted with people who have just voted. They are useful as they provide data at the very last minute, so that no other unforeseen circumstances can affect people's decisions (and undecided is no longer an option). However they are not available until the last minute, and can be more biased than opinion polls.
  • Margins of error: As censuses are expensive and ultimately pointless, given that the election is effectively a big census, with "did not vote" as an option, pundits (or "talking-heads") use surveys, which involves just interviewing a hopefully representative random sample of voters. This, however, means that the surveys results are not likely to be quite the same as an equivalent census. A margin of error reflects how much variation could likely be expected. Due to the fact that a sample was used, they do not cover issues such as a sample being unrepresentative.
  • Attack ads: Attack ads are a form of political campaigning where rather than emphasizing their own qualifications and attributes, a candidate or a group affiliated with the candidate tries to convince voters that their opponents are unsuitable for the office.
  • Game-changers: Game-changers are events that pundits claim will lead to significant changes in polls. During an election season, at least one election event a month for each candidate is promoted as a "game-changer" by the media. In actuality, it usually has no effect on the outcome, as most voters are already decided in US presidential elections.
  • Tracking polls: A tracking poll is a poll repeated at intervals and averaged. For example, a monthly tracking poll uses the data from the past month and discards older data.
  • Swing states: In the United States, several states have significant support for one party, and hence due to the fact that the electoral college gives all of a state's vote to a single candidate, small changes in votes do not change the party that has the majority of the vote, and hence does not change who receives that state's electoral college votes. Other states that have near equal support for each candidate are referred to as swing states, and they are particularly interesting for pundits, as relatively small changes in votes can have significant effects on the end result.
  • Swing votes: Swing votes are similar to swing states, expect that they refer to groups of people. If there is a significant social group, a large segment of which could possibly vote for either side, they are of particular interest for pundits.
  • Bradley effect: The Bradley effect is a theory that in elections containing a minority candidate, polls will often tend to overstate their level of support. It is theorized that this is due to people not wanting to appear racist whilst being surveyed, so some people who support a non-minority candidate may either claim to be undecided or plan to vote for the minority candidate. However, as voting is private, at that point they may reveal their true preference. The effect is named after Los Angeles Mayor Tom Bradley, an African-American who lost the 1982 California governor's race despite having appeared to have a significant lead in the opinion polls. This effect was particularly interesting for pundits in the 2008 US Presidential Election because of the running of Barack Obama, the first African-American presidential candidate. As there was no precedent for this at this level, some pundits were concerned about how much of his lead in the polls might be due to this effect.
  • <name> the <occupation>: A significant event in the 2008 US Presidential Election was a question proposed to the then Democratic nominee Barack Obama by Joe the Plumber. This promoted a variety of hitbacks and counter-hits. Cueball is referring to the tendency of the media at the time to refer to many critics of the two candidates using the snowclone "<name> the <occupation>," where <name> is replaced by the person's first name and <occupation> by their occupation (e.g. "Brad the Masturbator").

The title text is about statistician Nate Silver, who became something of a geek celebrity for his analysis during the campaign. He correctly predicted the outcomes of 49 of the 50 states in the 2008 election on his blog. It jokes that having him predict the outcomes of life decisions would make choosing the best thing to do very easy. So if Cueball ask Nate - "Should I sleep with her?", then Nate could give him a forecast like this: "Well, I'm showing a 35% chance it will end badly." Later, in 2016, Nate Silver's website, FiveThirtyEight, launched an advice column thus making the title text partially come true.


[Cueball sits at his computer desk, staring at his computer.]
Cueball: It's over.
Cueball: After twenty months it's finally over.
Cueball: I don't have to be an election junkie anymore.
[Close-up of Cueball's face and screen.]
Cueball: I don't have to care about opinion polls, exit polls, margins of error, attack ads, game-changers, tracking polls, swing states, swing votes, the Bradley effect, or <name> the <occupation>.
Cueball: I'm free.
[Cueball staring at his computer screen, full shot.]
[Cueball types on his computer.]
Tap Tap
[On screen:]
[The Google logo.]
'2012 polling statistics'

comment.png add a comment! ⋅ comment.png add a topic (use sparingly)! ⋅ Icons-mini-action refresh blue.gif refresh comments!


Is this genuinely a thing that Americans do during election years? Davidy²²[talk] 22:49, 10 November 2014 (UTC)

Yes 02:31, 14 April 2020 (UTC)
Oh my god yes. Actually, we start midway through the year before. 23:16, 23 July 2020 (UTC)
The first Democratic debate for the 2020 election was in June of 2019. Cwallenpoole (talk) 23:07, 30 November 2020 (UTC)
Absolutely. There's lots of coverage on the Presidential race for about 20-24 months before the general election, then a few months after that it's time to start thinking about midterms and off-year local elections. QoopyQoopy (talk) 00:56, 13 September 2020 (UTC)

I'm surprised no-one has pointed out yet how wrong Nate Silver was about the outcome of the 2016 election! -- The Cat Lady (talk) 21:16, 15 August 2021 (UTC)