Wikipedia:Polls are primary sources
Essay on editing Wikipedia
From Wikipedia, the free encyclopedia
On major issues or demographic blocks, there's often a wide variety of polling available that asks in-depth questions intended to gather information about them. Such polls are WP:PRIMARY sources and much be used with caution, since they can imply things that aren't stated in the polls. This doesn't necessarily mean they're unusable, but they should be used only in cases where the meaning of the poll is instantly and uncontroversially obvious to all readers. Even then, there may be WP:DUE weight concerns, since polls frequently ask many questions and hot-button issues are frequently polls; putting undue weight on a poll whose results are an outlier, in particular, may result in an unbalanced article.
 | This is an essay on WP:SOURCETYPES. It contains the advice or opinions of one or more Wikipedia contributors. This page is not an encyclopedia article or a Wikipedia policy, as it has not been reviewed by the community. |
 | This page in a nutshell: Polls are primary sources, and should be used with caution; they particularly shouldn't be used in ways that imply interpretations or which could be potential synthesis. |
Additionally, polls may sometimes produce WP:EXCEPTIONAL results. In those cases in particular it is important to have secondary coverage, both to verify its meaning and to establish that it is due as more than just an outlier; the absence of coverage for a poll being cited for exceptional statements in an article is a red flag that we're either drawing inappropriate conclusions or that the poll itself is not as significant as it seems.
Even when a poll can be cited directly, a lack of secondary coverage means that the weight due to it in an article is likely low; it should generally not be cited in the lead, nor should entire sections or paragraphs be devoted entirely to citations to primary polling.
The ideal way to use a poll is always to look for high-quality secondary coverage, preferably from sources with expertise in polling or the topic area being polled, which will indicate what the results mean and how much weight to give it. This essay is about citing a poll directly; when citing a poll via secondary coverage, most of the concerns here do not apply, and the few that remain come down to the reputation of the secondary source.
Things to avoid when citing polls
While primary sourcing to polls should always be used with some caution, there are some use-cases that are particularly common and worth avoiding:
- Polling with exceptional, unusual, or surprising results. Per WP:EXCEPTIONAL, such things require higher-quality sourcing, and may indicate that the poll is being used inappropriately; if a reputable primary source legitimately produces exceptional results, it will quickly attract secondary coverage.
- Citations that pull individual results from a larger poll to imply a result. Many polls ask large numbers of questions; if an editor goes through this and finds one question from deep within the results to highlight, they may be giving it undue weight or performing original research to support an unstated conclusion. Ask to see a secondary source to indicate that question's significance.
- Giving excessive weight to a single poll. It is rare for one primary source, with no secondary coverage, to be due more than a single brief sentence in the body at most. Putting it in the lead or devoting a section or paragraph to it in the body is often giving it undue weight.
- Combining multiple primary citations to polls to imply or add weight to a result. While doing this might seem like a solution to the WP:DUE issue above, it introduces more problems, since it could be a form of synthesis and original research; beyond this, the restriction on avoiding relying too heavily on primary sources applies to primary sources as a whole, so adding more primary sources makes things worse, not better. When multiple polls need to be cited, it is better to look for secondary sources that summarize and characterize them.