Hello Alejandro, I do my best with the relevant and free articles and abstracts, for my introduction.  After all our papers are mainly concerned about what we have done and discovered.  And I found that at times, peer reviewers who have more experience will hint or suggest a few more references.  Then, I will look up the refs, and add them.  All my papers were reviewed, whether thoroughly or less so; and all of them were improved by the peer review.  Thanks, have a good weekend.

Hello Alejandro, I do my best with the relevant and free articles and abstracts, for my introduction.  After all our papers are mainly concerned about what we have done and discovered.  And I found that at times, peer reviewers who have more experience will hint or suggest a few more references.  Then, I will look up the refs, and add them.  All my papers were reviewed, whether thoroughly or less so; and all of them were improved by the peer review.  Thanks, have a good weekend.

Alejandro -

When writing a paper in a rather specific area of mathematical statistics, there may be many articles which relate tangentially, but few if any which are directly related.  I think it good to provide a substantial number of references while making it fairly clear which ones are the most likely to be of interest, often by making references to specific pages.  I suppose this practice would be good for many other disciplines, at least in many situations. 

As for the authors, when writing on a topic for which there are few experts, you might know which ones you want to reference, but keeping an open mind for new contributors is important. 

When reading Miranda's response, I was reminded that I like to reference freely available literature on the internet, and provide a URL when possible.  A good reference is rather useless if unavailable.  For many years I could obtain help from research librarians, and I was particularly grateful for one available to me when I needed some old articles for a paper in honor of a well-known statistician's 80th birthday, but those librarians were basically all phased out by the time I retired. Everyone is cutting budgets, and authors are expected to find their resources online for the most part, it seems.  Textbook references are sometimes helpful. 

Anyway, established authors of reference papers can be of some use, but it is not a good idea to be too exclusive.  New contributors should be considered. 

I refereed an article for one of the most prestigious statistics journals not long ago, and I'm usually quite easy on authors, but I was appalled by the horrible job done by a group of "experts" on that paper, and soundly rejected it.   Usually statistics papers are done by one or two authors, but a panel of several authors, one with an extremely good reputation, did that paper, and they were unaware of information available for many years.  Their whole, uninformed approach seemed rather silly.  I'd say it was a very bad showing. 

Speaking of being a referee, that may often be a good way to become more aware of the literature, so you know what to reference, as well as being a useful way to contribute to society. 

For review-type papers, and for sections in papers that do contain a review, I think it may be particularly important to provide a number of references and/or a bibliography, for further reading in areas of interest to your reader.  

Cheers - Jim

Miranda and James,

I coined this question for the sake of a discussion with grad students (in their first steps as paper-writers), so they can see what other people, besides their proffesor/tutor, think on the matter. I formulated the question having in mind that many students and young researchers might be taking the exchange as a lesson. ...and, what a lesson they are receiving! Thank you very much to both of you for sharing with me so great answers.

Hopefully, more people will share their opinions too, and we´ll try to integrate/analize them together.

Abrazos - Ale

Pick articles that are most focused to the statement you wish to support...and put more emphasis on recent work.  

At least half of the articles cited in your literature review should be from within the last five years, to demonstrate that you are basing YOUR study on the most recent literature.

I choose articles published in high quality journals

Best Regards Alejandro Bortolus