The Good article criteria are the requirements that an article should meet in order to be listed as a Good article (GA). Any articles that have been reviewed, and meet these criteria, should be listed as GAs. Good articles are "satisfactory" or "decent" articles, not great articles. The standards for GAs are fairly high, but noticeably lower than the Featured article criteria. Importantly, the GA criteria are a standard, not the opinion of individual reviewers. This essay is intended to help reviewers make their decisions based on whether an article meets the criteria or not, rather than personal preferences.

Article editors and reviewers should have as a common goal the ideal to make the article as good as it can be. However, the decision to list or not list an article should be based on the GA criteria alone. Reviewers are encouraged to differentiate clearly between those improvements that are necessary for the article to meet the criteria, and suggestions to improve the article beyond the actual criteria. As a rule of thumb, if none of the six criteria plainly apply to a suggested improvement, it should be labeled as such or considered optional.

Regular editors of a nominated article should likewise assume and act in accordance with good faith and the goal of article improvement. The reviewer may have less expertise in the subject matter. If they make impractical or inappropriate suggestions (e.g., for the inclusion of information that isn't known, or images that simply don't exist), explain politely why this is impractical or inappropriate. Regular reviewers often have good suggestions for improvements that go beyond the GA criteria. These suggestions are optional with respect to GA status, but implementing them may result in an even better article, which may help it reach Featured article quality.

If editors and reviewers reach an impasse, Community reassessment is available to resolve disagreements, but all are encouraged to resolve as many outstanding problems as possible first.

A good article is—

Point (a) means:

• The meaning of each sentence or paragraph is clear and not confusing, even if you might have phrased it differently.
• The spelling and grammar follow an established system, even if you prefer a different variety of English.

Point (b) means:

• The five particular pages listed (out of about 50) of the Manual of Style are complied with. (The "Fiction" and "List incorporation" guidelines do not apply to many articles.) Note that the main Manual of Style page is not in the list of required MoS pages. It may, however, be a useful page to look at if you have questions about spelling and grammar for point (a).
• For the purpose of a GA review, all other parts of the MoS are optional except captions and image relevance, which are discussed in part (6) of the GA criteria.

Mistakes to avoid

• Imposing your own stylistic preferences or national variety of English on the article text.
• Demanding compliance with your favorite MoS pages.
• Requiring the elimination of all bulleted lists, even if they are acceptable under WP:EMBED § Appropriate use of lists.
• Requiring the resolution of links to disambiguation pages. The fact that the toolbox provides a handy tool for checking these things doesn't mean that the links are prohibited by the GA criteria.
• Requiring the removal of red links. (Even FA and FL permit red links.)
• Requiring common sections (like cast lists) to follow a particular format or to look like similar sections in other articles.

Point (a) means that there must be at least one section with a ==Level 2== header, containing a list of sources used in the article. That section heading should have a title that identifies the contents that the reader will find there, e.g., ==Notes== or ==References== or ==Sources cited== or ==Footnotes== or anything with a similar meaning. It is extraordinarily unusual for a nominated article to not pass the 2(a) criteria, and if you think you have encountered one, then you should seek assistance at Wikipedia talk:Good article nominations.

Point (b) names five types of statements for which the good article criteria require some form of inline citation:

1. direct quotations,
2. statistics,
3. published opinion,
4. counter-intuitive or controversial statements that are challenged or likely to be challenged, and
5. contentious material relating to living persons.

This standard is higher than the absolute minimum standard set by policy, but noticeably lower than many editors' personal preferences. If an article contains none of these five types of statements, then Wikipedia:Citing sources § General references may be used. If the article contains any of these five types of statements, then those specific statements must carry inline citations which allow the reader to connect a specific sentence to text in some source. Other article text may be supported by inline references or general references.

Page numbers (or similar details) are only needed when the inline citation concerns one of the above five types of statement and it would be difficult for the reader to find the location in the source without a page number (or similar detail).

Point (c) means that all facts, opinions and synthesis in a good article should be based on reliable sources with no original research. Statements made in the article should reflect the material in the sources.

Point (d) means that none of the text is copied from another work without proper attribution. All of the text is either properly paraphrased or quoted and is sourced accordingly. If the entire article is copied from a non-free source, not only does it fail this criterion, it makes it eligible for speedy deletion under criterion G12.

Mistakes to avoid

• Imposing personal preference on reference section headings.
• Asking for inline citations where they are not necessary, such as in plot summaries or content which summarizes cited content elsewhere in the article.
• Not checking at least a substantial proportion of sources to make sure that they actually support the statements they're purported to support. (Sources should not be "accepted in good faith": for example, nominators may themselves have left material added by prior editors unchecked.)
• Requiring page numbers where these are not essential.
• Demanding the removal of dead links, in direct violation of WP:Linkrot and WP:DEADREF
• Requiring the use (or non-use) of citation templates.
• Requiring consistently formatted, complete bibliographic citations. If you are able to figure out what the source is, that's a good enough citation for GA.
• Requiring consistent date formatting.
• Requiring that footnotes be listed in numeric order, if multiple citations are named after a sentence.
• Rejecting reliable sources because they are in a language you don't happen to read. See Wikipedia:Translators available for potential assistance.
• Not checking if the article is committing plagiarism or violating copyright.

Point (a) means that the "main aspects" of the topic, according to reliable sources, should each be "addressed" in the article; it does not require comprehensive coverage of these major aspects, nor any coverage of minor aspects. For particular types of article, WikiProjects often provide helpful advice on what the main aspects are likely to be. (Do not, however, require compliance with any advice pages written by WikiProjects; only compliance with the policies and guidelines specifically named in the six Good article criteria themselves are required.) For an article on a work of fiction, a summary of the plot and a discussion of the reception are usually required. For an article on a disease, the causes, symptoms and treatments are usually significant.

Point (b) raises two issues. First, the article should avoid undue emphasis on tangents, such as coatracks, and trivia. The inclusion of details and minor aspects can contribute to good writing, but such details should not overwhelm the article. Second, the level of detail of each aspect of the topic should be appropriate to the article and kept in balance: where an aspect of the topic involves information which is or could be covered in more detail by another article, the article itself should summarize this information with suitable links, such as {{Main|(the background article)}}, where appropriate.

Taken together, these criteria mean that no obviously important information should be entirely absent from the article, and the level of detail should be appropriate to the significance of the information. It is better to have an article that covers the essentials well, based on reliable sources, than a diffuse article relying on trivia or unreliable sources to flesh it out.

These criteria do not impose arbitrary size restrictions (in terms of kilobytes, characters or readable prose). Good articles can be as short or long as is appropriate to the topic: WP:SIZE is not a good article criterion. However, size issues may be indicative of genuine GA problems with coverage (3a), concision and focus (1a and 3b), or the use of summary style.

Mistakes to avoid

• Requiring lengthy sections, especially if the same information can be adequately presented in a concise form.
• Requiring the inclusion of information that is not known or addressed by reliable sources.
• Not noticing that a major aspect is completely omitted from the article, despite being discussed significantly in reliable sources.
• Imposing arbitrary size restrictions, rather than directly addressing GA issues of coverage, conciseness, focus and the use of summary style.

This point requires that the article is presented from Wikipedia's neutral point of view. The subject should be discussed in encyclopedic language, based on reliable sources, with due weight given to all significant viewpoints.

For articles on uncontroversial topics (such as a species of plant or animal), this is likely already covered by criteria 1, 2 and 3, especially if you watch out for word choice issues (1b). For controversial subjects (biographical, political, religious and health articles require particular care), you need to be sure that significant viewpoints are fairly and accurately represented, based on the diversity of reliable sources available (not just those favoring one viewpoint), and that the article does not endorse or favor a particular view through word choice, sentence structure, section titles, or article organisation. An article written from the neutral point of view provides the reader with information, allowing them to form their own conclusions.

Mistakes to avoid

• Requiring excessive representation of minor or insignificant viewpoints.
• Requiring that all viewpoints be presented as equally valid.
• Listing as GA a controversial article written from a particular viewpoint, or which leads the reader to a particular conclusion.

The footnote here is important: "Vandalism reversions, proposals to split or merge content, good faith improvements to the page (such as copy editing), and changes based on reviewers' suggestions do not apply."

Good article reviews are not supposed to interfere with normal editing. An article is unstable if there is a significant edit war underway, if editors are directly telling you that you shouldn't review the article because they're in the middle of major changes, or if the article is changing so dramatically and so rapidly that you can't figure out what you're supposed to be reviewing.

Mistakes to avoid

• Discouraging normal editing activity for the convenience of the review.
• Confusing normal talk page discussions, incremental improvements to the article, or vandalism, with a real content dispute or an edit war.

Most GAs contain at least one image (or other media: video, sound clip, etc.), but suitable images may not be available.

If images have not been included, and if suitable images are not readily available (checking Commons for images is a good idea), then this criterion is automatically satisfied. If you think that free or fair-use images should be readily available, then please either find and add the images yourself, or recommend specific sources or images to editors.

If images have been included:

Point (a) requires reviewers to click every image (sound clip, etc.) to check its copyright status. If it is a free image (i.e., is in the public domain, or is released under a free license such as GFDL or Creative Commons) then Point (a) is satisfied. If it's a non-free image, then it must have a valid fair use rationale that specifically justifies its use in the article under review.

Point (b) says that every included image must be relevant to the topic, and must have a suitable caption. Purely decorative images, such as an image of a butterfly in a psychology article about emotions, should be removed. WP:ALT text, although easy to provide, is not required.

Mistakes to avoid

• Failing the article because no free images or other media currently exist.
• Making vague requests for "more" or "better" images.
• Accepting or requiring images which are decorative but irrelevant.
• Requiring compliance with all of MOS:IMAGES (as opposed to just the sections linked in the GA criteria).

The good article criteria and good article processes are designed to address article content quality. Although there are overlaps, the following issues are beyond the scope of the criteria.

• Notability issues, such as whether the subject merits its own article: questions related to notability and deletion are discussed at Articles for Deletion. Content forking and merging issues are also not within the remit of good article processes. If you see a nomination whose notability you personally doubt, please do not start a GA review yourself. Instead, nominate it for deletion. If the article is kept, then let someone else review it.
• Article titles are not addressed by the good article criteria. These are discussed at Requested Moves.
• Legal issues (other than copyright violations), such as those associated with biographies of living persons. These should be raised at a relevant forum, such as the BLP noticeboard.
• Content disputes. Although the good article criteria require article stability, good article processes should not be used to further content disputes, nor are the criteria designed for dispute resolution. Use the article talk page instead, and seek mediation if necessary.
• Compliance with the External links guideline is not required. Feel free to tag a {{linkfarm}} whenever you see one, but it is not usually appropriate to consider the contents of the External links, Further reading, or See also sections when deciding whether the article meets the GA criteria, because these sections are not mentioned in the GA criteria.
• Infoboxes and navigation templates. The GA criteria neither require nor prohibit the inclusion of these, so their presence or absence should not, in itself, affect the review. However, as with all material in an article, the information in an infobox (such as titles, definitions, or statistics) is required to meet the GA criteria. Since information in an infobox is usually repeated in the body of the article, this is rarely a concern.
• Article assessments by WikiProjects for the WP:1.0 team. The assessments at the top of article talk pages are frequently reviewed less than once a year. They should be assumed to be out of date. At the end of your review, you have no obligation to update the WikiProjects' ratings if an article is listed as GA, but it is helpful (just change |class= from whatever it was to |class=GA). The previous rating should not influence your evaluation of the article at all.
• Tedious style nit-picks, beyond the points in the specific MoS pages incorporated into the GA criteria. There are many MoS pages on various things. A GA needs only to comply with the basics, though fuller compliance will be expected for featured article candidacy.
  • Nominators may wish to do an MoS-compliance editing pass before a GA nomination, just to avoid a GA review becoming bogged down in stylistic quibbles. WP:WikiProject Guild of Copy Editors may be willing to help, especially if you tell them this is in preparation for GA. Reviewers will usually find that it takes less time and effort to just do simple MoS cleanup themselves, than to list out a bunch of desired cleanup and an explanation for each point in a review: Why spend time typing something like "In the third paragraph of ==Early life and education== section, the last sentence has an obvious typo in the word the" when you can just fix it in five seconds? (Reviewers are permitted to fix small problems they find in the articles they are reviewing.)