Wikipedia:Five percent rule

Several RFCs have established that third-party candidates must poll over five percent to be included in an infobox. The only exception is if only one candidate polls over five percent, in which case the second-place finisher may be included if determined appropriate by local consensus. This does not strictly apply to parliamentary elections, where other criteria may be used to include candidates if deemed appropriate.

Origin

The notion of a "five percent rule" appears to originate with Walter Dean Burnham's 1970 essay Critical Elections and the Mainsprings of American Politics. Page 4 of Three's a Crowd: The Dynamic of Third Parties, Ross Perot, and Republican Resurgence states "most scholars follow Walter Dean Burnham, who defined 'successful' third parties as those that attract at least 5 percent of the vote. (Burnham 1970)".

Summary of discussions

Link	Result	Notes
2010 discussion	(Apparently, a rough consensus for a 5% threshold)
2013 RfC	"There was a standing consensus, mentioned by many below (though I'm not sure it ever made its way into guidance somewhere) that a 5% threshold was sufficient. In the linked discussion, many 3rd party sources were put forth explaining why 5% was meaningful, and not arbitrary. The bulk of the !no votes felt nonetheless that 5% was arbitrary, but this goes against the past consensus on this issue, while the yes !votes provided more detail on polling numbers and 3rd party mentions in RS of this candidate. This discussion has nonetheless highlighted a need to have a broader discussion on the topic of when a candidate deserves to be in an infobox, so I suggest a new discussion be opened at Wikipedia_talk:WikiProject_Elections_and_Referendums to propose criteria for inclusion in a general sense, and I encourage people to start that discussion and focus on clear, crisp criteria to the maximum extent possible (It is fine if the criteria are specific for the United States, and eds can come to consensus for other countries as needed. Once that discussion is concluded and further consensus is reached and added to a guideline somewhere, this particular issue can be immediately revisited, but for now my judgement is that there is a rough consensus based on strength of argument and WP:NPOV that argues for inclusion of the candidate. If per newly decided consensus criteria he no longer qualifies, then at that point he could be removed."
2014 RfC	"Robert Sarvis should be included in the infobox. Two editors felt that Sarvis should be excluded from the infobox and five felt that he should be included. One editor appeared to be neutral/leaning exclude. Since there doesn't appear to be any clear dividing line between where a candidate should be included in an infobox and where they shouldn't (given that not all elections are alike, perhaps it is right that there should be no clear dividing line), I don't see what else can be done here except to go with the majority. It was claimed during the discussion that there is an existing consensus that candidates must get 5% or more in at least two opinion polls to qualify for the infobox, but I have not been able to find any evidence of this. If evidence is presented after the close, then I may be willing to take it into account."	Editors were unaware of 2013 RfC
2016-2017 RfC	"There is consensus that Donald Trump should be included in the infobox despite achieving less than 5% of the vote per WP:IAR. The 5% benchmark was established in a small discussion, and was intended to be applied to third party candidates. A follow-up discussion addressing the general criteria for inclusion in the infobox is encouraged. 5% of the vote or a major party nomination was floated below, and gained some support."	Editors were unaware of 2013 RfC, and only cited the short 2010 discussion.
2017 RfC	"Result: The standard for inclusion in the infobox of U.S. election articles is: A candidate must receive 5% of the vote. If only one candidate meets the threshold, then the second-place contender should be included. Question 1: There was minimal support for option B, a 4% threshold, rendering it non-viable. The debate essentially reduces to a question of whether or not to add "or be the nominee of a major party" clause to the current common practice 5% threshold. There was more support for A, a simple 5% rule. For practical purposes the non-viable B votes appear to translate into A votes. However there was still solid support for C, adding a major party clause. Given the fuzzy nature of treating B votes as A votes, and the fact that this still leaves the result fairly close, I find it most appropriate to call this "No consensus to add a major party clause". This has the practical effect of defaulting back to A, a basic 5% threshold. Question 2: There was strong consensus for A, "If only one candidate meets the threshold, then the second-place contender should be included"."
2020 RfC	"This is an interesting and well-designed RfC concerning infoboxes. Generally, a Wikipedia article should be a summary of the reliable sources, and where editors have chosen to use an infobox, the infobox is a summary of the Wikipedia article. The basic job of an encyclopaedist is to write summaries, and Wikipedians generally have a lot of practice at it, so we don't normally find infoboxes difficult. However, not everything fits neatly into a pithy entry in an infobox, and that's why infoboxes come up so much at RfC. In the matter of elections, previous discussions have reached consensus that if the article contains an infobox, then the infobox should usually contain at least two candidates, even in cases where the winning candidate received a very overwhelming share of the vote. Like all content decisions, this consensus allows for occasional exceptions in unusual cases. This discussion seeks to establish numerical thresholds for the inclusion of second place candidates. Wikipedians have not been able to agree on any such set of thresholds, so in some cases decisions will need to be made individually on article talk pages. However, the 5% threshold does enjoy plurality support -- even if this support falls short of an actual consensus -- so I suggest that the proposed 5% threshold could be a useful reference point in these talk-page discussions. I hope this helps"
2021 RfC	"I find consensus in support of this proposal, with the caveats that 1) different principles may apply to parliamentary/party-list elections and 2) as always, appropriate exceptions may be made via local consensus."	Established that the 5% rule (and the exception for second-place finishers below 5%) applied to all elections, American or not.
2023 RfC	"Consensus was reached that [the 5% rule] should not be strictly implemented for parliamentary elections."
2024 RfC	"This discussion was complicated and involved evaluating many overlapping criterion. This discussion also got auto-archived by the bots, which further started another discussion, which was also considered when deciding this consensus. Overall, there is clear consensus in favour of Criterion 6 and Criterion 1a being applied to this article. Before the elections, candidates can be included in the infobox if they have ballot access in states that comprise 270 electoral votes and generally polls at 5% or above in major polling aggregators. (RealClearPolitics, FiveThirtyEight, et al.). A supermajority of editors preferred or were okay with both the criterion stated here. While some editors argued in favour of using reliable sources' as an alternate barometer for "seriousness" of candidacies, nothing precludes us from using polling directly as such a threshold. The subsequent discussion below resulted in similar positions and a proposed compromise, resulting in more variety of final positions. Overall, said discussion involved fewer editors and did not have a strong enough agreement to overturn the original RFC consensus. There is precedent for a similar threshold via Wikipedia:Five percent rule, although some editors disputed whether or not it applies to pre-election polling. Regardless, there is a clear local consensus established here in favour of Criterion 6 and 1a."	This RFC explicitly indicated that the "results of this RFC should not be interpreted as WP: PRECEDENT outside of this article"