As the June 26 Dispatch mentioned, generally the most reliable sources in medicine and biology are peer-reviewed publications in scientific journals. An excellent starting point for locating peer-reviewed articles is PubMed, which offers a free search engine for accessing the MEDLINE database of biomedical research articles offered by the United States National Library of Medicine at the National Institutes of Health. Although PubMed is a comprehensive database, many of the journals it lists restrict online access; at the alternative site, PubMed Central, access to the full text of articles is free of charge.

There are two main types of sources in the scientific literature: primary publications, which are papers describing novel research for the first time, and review articles, which summarize and integrate a topic of research into an overall view. In medicine, there are also clinical trials, which test new treatments, and meta-analyses that bring together the results from many clinical trials and attempt to arrive at an overall view of how well a treatment works. It is usually best to use reviews and meta-analyses where possible, as these give a balanced and general perspective of a topic, and are usually easier to understand!

The quality of sources varies considerably, and assessing it is difficult. To know how much weight to give a particular publication usually requires a good grasp of what has been published and how this fits into other people's results – this is why reviews are so useful, since they do this for you. However, two rules of thumb can be useful. First, if a biology/medicine journal is not listed in PubMed, it is of doubtful quality – the journal published by the Creation Research Society is one example. Second, the impact factor of a journal, published yearly in Journal Citation Reports, can tell you how influential it is. This number measures how often papers in the journal are cited by other papers – good journals publish papers that other people find useful, while bad work sinks into well-deserved obscurity. These impact factors are not the definitive word on reliability and vary significantly from field to field; however, occasionally they may assist editors in knowing how seriously a source will be regarded by expert readers.

Other than peer-reviewed journals, the most reliable sources are conference proceedings and books published by university presses and other respected publishing houses. Wikipedia also accepts sourcing from non-peer-reviewed academic sources, such as press releases from respected universities, government research organizations, scientific societies, and from reliable non-academic publications, particularly if they are respected mainstream publications; these include professional and trade magazines, journals, and newsletters. Mainstream newspapers may contain reliable information, but it is often wise to check the reliability and status of particular journalists. A non-peer-reviewed source is sometimes also cited because it is a readable lay summary of a peer-reviewed article; here is an example:

Johnson CP; Myers SM (2007). "Identification and evaluation of children with autism spectrum disorders". Pediatrics. 120 (5): 1183–215. doi:10.1542/peds.2007-2361. PMID 17967920. See also this lay summary.

Other things being equal, it is better to cite a source whose full text is freely readable, so that your readers can follow the link to the source. However, many high-impact journals, such as Nature and Science, require a fee or a subscription, and as these journals publish some of the best papers it can still be best to cite them.

Often an article's abstract – usually 100–200 words summarizing the findings and significance – is freely available if its full text is not. When searching for sources, it is wise to skim-read everything you can, including abstracts of papers you can't access, and use that to get a feel for what reliable sources are saying. When it comes to actually writing your Wikipedia article, though, it is generally not a good idea to cite a source after reading only its abstract, as the abstract necessarily presents a stripped-down version of the conclusions and omits the background that can be crucial for understanding exactly what the source says. You may need to visit a library to access the full text, or ask somebody at Wikipedia:WikiProject Resource Exchange to either provide you with a copy or read the source for you and summarize what it says; if neither is possible, you may need to settle for using a lower-impact source or even just an abstract.

Some sources are in the public domain. These include many U.S. government publications, such as the Morbidity and Mortality Weekly Report of the Centers for Disease Control and Prevention. You can incorporate public-domain text into a Wikipedia article without infringing copyright, which can help you write an article on a new topic quickly. However, in such cases you should follow scholarly practice and cite the source, putting quotation marks around direct quotations, or using the blockquote facility for longer quotations. The amount of directly quoted text can sometimes be minimized by working it into the grammar of a Wikipedia sentence. It is often better to paraphrase a source (without quotation marks but with attribution), which allows you to summarize it, improve the wording, and turn it into an encyclopedic register.

Each search engine has its quirks, advantages, and disadvantages, and may not return the results that you need unless used carefully. It typically takes experience and practice to recognize when a search has not been effective; even if you find useful sources, you may have missed other sources that would have been more useful, or you may generate pages and pages of less-than-useful material. A good strategy for avoiding sole reliance on search engines is to find a few recent high-quality sources and follow their citations to see what your search engine missed. It can also be helpful to perform a plain web search rather than one of scholarly articles only.

There are basic and advanced options for searching PubMed. For example, if you enter keywords such as "breast cancer" in the basic option, you'll be flooded with a six-figure number of results. At the top right of the list of hits there are three links, one labeled "All", the second "Review" and the third "Free Full Text". Clicking on the "Review" link will slash your hits to a more manageable 14,000 review articles.

To look at one of these in more detail, just click on the title – such as RE Coleman's review, Risks and benefits of bisphosphonates. This takes you to a summary (the abstract) of the review, providing a list of authors and, on the right, a list of related articles. This "See all Related Articles" link is very useful for narrowing down searches. At top-right there can be a link to the journal website (here it is the British Journal of Cancer). At the bottom of the abstract is a number called the PubMed ID number, which is "PMID 18506174" in this instance. To generate a {{cite journal}} template for Coleman's review, just copy the PMID number into the Diberri-Boghog tool.

To perform an advanced search with the same keywords, go back to the search screen, enter "breast cancer", as before, and then click on the "Advanced Search" link just above the search box. This takes you to a set of options that allow you to limit your search to particular dates, types of article, and topic areas. For example, to search for meta-analyses, tick that box in the section on Type of Article and hit "search". This will generate a list of about 460 meta-analyses that deal with breast cancer.

If your Pubmed search finds a lot of sources, you can restrict yourself to the freely readable ones by clicking on the "Advanced Search" link and checking the box labeled "Links to free full text".

Pubmed is not the only game in town for biomedical searching. There are alternatives.

• Some search engines attempt to cover all scholarly sources. They are invaluable for topics not covered by the more specialized indexes, and can provide useful sanity checks even for topics such as medicine that have more specialized indexes. The best-known is Google Scholar; other engines include getCited and Scirus. A version of Google Scholar that automates the generation of citation templates is WikiScholar.

• The Cochrane Library contains a database of systematic reviews and meta-analyses, and is a key resource in evidence-based medicine. Its reviews are generally considered to be of very high quality.

• EMBASE is a high-quality index that often generates better results than Pubmed. Unfortunately, it is proprietary and requires a subscription.

• CINAHL is a proprietary index on nursing and allied health care.

• To be indexed in Pubmed, journals must meet editorial and content quality standards. When writing articles on non-mainstream topics, it may be useful to search indexes that include journals that do not meet these standards. You must be careful not to misrepresent these publications as reflecting mainstream opinion. These indexes include AMED for allied professions and complementary medicine, Alt HealthWatch for complementary and alternative medicine, ICL for chiropractic, and MANTIS for manual medicine.