If you need help while reviewing, you can request assistance through reviewer mentorship.
Reviewing a good article nomination (GAN) means reading it carefully to make sure it meets the good article criteria then officially informing the Wikipedia community that it passes or fails GA. It's a great way to help out with an article and to improve your own editing skills in the process. If you have an account, then you can review any good article nomination as long as you haven't significantly contributed to the article. You don't even need to be knowledgeable about the subject. Find a nominated article that isn't already being reviewed by someone and review it.
Starting the review
First, make sure you read the good article criteria. Keep them in mind while you're reviewing the article. Then go to Wikipedia:Good article nominations and pick an article to review. Once you've picked the nomination you want to review, click where it says start review and save the page. If you'd like, you can insert a GA template to decide on a format for your review. Check that the nominator is actually a major contributor to the article and that they didn't nominate an article written by someone else.
Once you save the page, you are that article's reviewer. Make sure you complete the review in a timely manner, because the nominator will be looking forward to hearing your feedback. Reviewers usually finish their reviews in a few days to a week. If you expect it to take longer, try to let the nominator know ahead of time.
You can immediately end the review as a fail if:
It is a long way from meeting any one of the six good article criteria
It has (or it should have) obviously valid cleanup banners like {{cleanup}}, {{POV}}, or {{unreferenced}}.
It was re-nominated after failing a previous review without fixing obviously valid issues that had been raised.
Reviewing the article
Read the whole article carefully, including the sources section. Check to see if it meets the good article criteria. Whenever you find an issue, make a note of it and explain what needs to be fixed. Use these guides to check the article:
1. Well-written
Clear and concise
While reading the article, check to make sure that it's readable and that it makes sense. List each way that the article is not clear, is not concise, is not readable, or does not make sense. Be specific about what issues need to be fixed, but do not demand that the nominator do a full copyedit beyond what is listed below:
Is it clear what each sentence is trying to say? Or, are sentences vague or confusing?
Is it concise? Or, do sentences meander on or use too many words to say something simple?
If the subject is fictional, this should be made clear in the first sentence of the article.
If it's about a work of fiction, it must avoid in-universe perspective (acting like the events of the story actually happened).
When describing what happens in a fictional work, the article should use present tense.
Plot summaries should be brief (usually under 400 words for TV episodes, under 700 words for other works). They should not give undue weight to any one event in the story.
The plot summary must not interpret or analyze the story and its meaning. It should just describe the sequence of events as they appear to the audience.
If you find simple mistakes like misspelled words, missing punctuation, or things that should be linked, you should be bold and fix them yourself. For anything more complicated, just mention it in your review and the nominator will fix it.
2. Verifiable with no original research
Check for verifiability, in other words, attributable to reliable, published sources. Do each check below, and make a note that you checked each one in the review. Make a list of each sourcing or citation issue that you find.
Check to make sure that all of the article's references are listed correctly in a section labeled "references", "works cited", etc. Any reference formatting is fine; your only concern is whether you can tell what the source is. There should not be any bare URLs; the reference needs to say what it is and where the link leads.
Look at the references and make sure that they are all reliable. Wikipedia:Reliable sources/Perennial can help you with the most common sources, but you may have to use your judgement. Check to see if the sources are written by authors who are experts in that subject and if they're published by reputable organizations. The best sources are up-to-date secondary sources that reflect the mainstream consensus about a topic. The worst sources are blogs, user-generated content, and self-published sources; these are almost never acceptable.
Check that the article uses inline citations. Not every sentence needs to be cited, but most claims should have a citation that supports it. Leads, plot summaries, and image captions don't usually need citations. Quotations, statistics, opinions, controversial claims, and contentious material about living people must have an inline citation no later than the end of the sentence, no matter where they are.
Check that there aren't too many quotes in the article. A few quotes are fine when they're relevant in context, but too many quotes is a type of plagiarism. If there are too many quotes, mention in the review that some of them need to be paraphrased.
Tools like WP:EARWIG and WP:CITEUNSEEN are helpful when analyzing an article's use of sources, but they only catch the most obvious problems. You'll need to check the sources individually to confirm that they're acceptable.
Spot checks
To review an article, you need to access its sources and compare them to what it says in the article. This step, called spot checks, is really important, because it checks whether the information in the article is true and that the wording is not stolen from somewhere else. Pick a few sources from the references list to check. 10% of them is a good rule of thumb, but try to check at least five of them. The sources will usually have links where you can access them. If you're finding it difficult to access most or all of the sources, ask the nominator if there's a way to access them or if they can provide you with the relevant excerpt.
Each time you access a new source, follow these steps:
Read the source, compare it to what is stated in the article, and make sure the article actually says what it's supposed to.
Check to make sure the source wasn't just copied-and-pasted into the article. If the Wikipedia article is identical or nearly identical to the source, it's a copyright violation.
Check for close paraphrasing. This is when an article isn't directly copied from the source, but when it presents information in a very similar way as its sources, using similar wording or describing things in a similar way. If you can match sentences from the Wikipedia article to the source one-to-one, that's evidence of close paraphrasing and you should note it in your review.
Check if the article is claiming something that is not stated in the referenced source. That would be a sign of original research.
List the source that you checked, and note whether you were able to confirm the information and whether you found any of these issues.
If some of your source checks turn up problems, check a few more to see if it's a significant problem throughout the article. Again, note each source you check and list any issues you find.
If you find that a significant proportion of the article is copied-and-pasted from other places, then quick fail the review and list the article as a copyright violation at Wikipedia:Copyright problems.
Remember that sometimes other websites might copy from Wikipedia. Check to see if that's what happened before saying something is a copyright violation.
3. Broad in its coverage
Make sure the article addresses each major aspect of the topic. It doesn't need to be totally comprehensive, but all of the most important aspects should be covered. For a biography, this would mean that every period of the person's life should be given some coverage. For a film, this would mean some coverage of the plot, cast, production, release, reception, and themes. An important section should not be missing. If you're not sure what aspects should be covered, research the subject to see what reliable sources say about it, or look at other articles about similar topics to see what's covered there. If sources don't cover a certain aspect, then the article probably doesn't need to either.
Also make sure that the article doesn't go into excessive detail. Any unhelpful trivia should be removed. Anything that is not part of the article's scope should also be removed.
Opinions or controversial statements should not be presented as facts. Any time an opinion is described, an article should make it clear whose opinion it is.
Agreed upon facts should not be presented as opinions.
Language in the article should not suggest or imply that something is good or bad.
Wikipedia articles should describe arguments, not participate in them. If you can figure out the author's opinions just by reading the article, then it needs to be rewritten.
Contentious labels must be cited to high quality sources, especially if they're about living people.
The article should also be neutral in its structure. This means:
Fringe views and other views in the minority should not be given the same weight as major views. If most reliable sources accept one position, then alternatives should not be presented as if they are prominent.
Minor aspects of a topic should not be given undue importance. Even if they're not controversial, it's still a neutrality issue.
If there are competing views of similar prominence in reliable sources, then one should not be given additional weight over the others.
No individual source should have outsized influence on the article, nor should the ideas of one specific writer, expert, critic, or school of thought.
The article's purpose should not be to take a specific stance on a topic. This would be a POV fork, which is not allowed on Wikipedia.
Neutral point of view is most often a problem with politics and other contentious topics, but it can happen anywhere. Always check to make sure the ideas in the article reflect the general consensus in reliable sources.
5. Stable
Check to make sure the article is stable. Check the article's history and its talk page to make sure there aren't any ongoing disputes surrounding the article. Constructive editing and collaboration do not count against an article's stability. Nor does vandalism. If there is a dispute, hold or fail the article so there's time for the dispute to end.
6. Illustrated
A good article should have media files, such as images, video, or audio, although it is understandable if this is not possible.
Once you've gone through all six criteria, you should have a list of everything you've checked and every issue that needs to be addressed. Post this list to the review page. Based on how close the article is to the good article criteria, there are a few things you can do:
Pass – If the article already meets all six of the good article criteria, you can pass it immediately after completing your review. This is very rare, as nobody is perfect and nearly all articles will have at least one thing to fix.
Fail – If you've discovered serious problems with the article while reviewing it and you anticipate it will take a long time to fix it, then you can fail the article.
Hold – In most cases, you will hold the article. This means that you're giving the nominator time to look at your review and fix all of the issues that you found.
Second opinion – If you need another reviewer to join the review for a second opinion, you can request this. This is usually only necessary if there's an unusual circumstance or if there's a disagreement between the reviewer and the nominator. Other opinions can also be sought more informally by posting at the good article discussion page.
If you need help with these actions, the GANReviewTool can do them automatically.
Assuming you put the article on hold, the nominator will be notified and they should begin addressing any issues with the article. Keep an eye on the review during this time, as the nominator may ask for clarification or give input on certain issues. Once the nominator says that they've fixed all of the issues, check to see if the fixes are sufficient. If all of the issues are fixed, then you can pass the article and the review is complete. Otherwise, note any issues that still exist.
Nominators will typically respond to all requests within a reasonable timeframe. How long this takes may vary, but seven days is a common estimate. If the nominator does not address the issues and you are unable to get a response from them on the review page after a reasonable amount of time, then you should fail the article.
Common mistakes
Checklist reviewing – Don't just mark the criteria as pass or fail without providing feedback. If the article passes a criterion, why? What did you check for? If it fails a criterion, what needs to be improved? Unless you're quick failing the article, you should provide a list of each item that needs improvement. If you are quick failing the article, you should still provide a summary of what the major issues are and what needs to be fixed.
Excessive standards – An article doesn't have to be perfect to pass, it just needs to meet the good article criteria. Don't count it against an article if it's not written the way you would have written it, if it doesn't follow a certain format, if it's still short after covering the important information, or if it has any other imperfections that are not addressed by the good article criteria. Offering additional suggestions for improvement is always acceptable, but make it clear to the nominator that they are just suggestions and they are not part of the good article criteria. Don't demand that nominators remove red links, mimic the structure of similar articles, use a certain date format, etc.
Prioritizing minor issues – Things like punctuation and spelling can be quickly addressed, and it's time consuming to list every single instance on a review. If it's a minor issue, the reviewer should fix it where they see it. If it's a problem throughout the article, then the problem can be stated without listing each example.
Formatting sources – Don't confuse the aesthetics of the sources with their content. All that matters is that it's verifiable. If you can tell what the source is, then there is no reason to ask the nominator to alter its formatting. When evaluating sources, you should be checking what is in the book/website/newspaper/etc, not what the footnotes look like.
Rapid fire review grabbing – Avoid starting several reviews at the same time. This is unfair to the people who have to wait for you to get through them, and it's unfair to yourself. Try not to take on additional reviews until you've finished most of the work on any previous reviews you're working on.
Resources
Most issues can be solved by speaking with the nominator. If you need help with your review or there's a dispute that can't be resolved, consider requesting a second opinion. If you have another issue that needs to be resolved, there are many places on Wikipedia where other editors are willing to help out:
Wikipedia:Good article mentorship – If you need more detailed help, you can request a mentor here. A mentor can help you with any issues you might have when reviewing.
Wikipedia:Neutral point of view/Noticeboard – If there's disagreement about how much weight a certain idea should be given in the article or about any other neutrality issues.