Finding and Reading Research Papers

By the end of this guide you will be able to find the research papers that matter for your question, get the full text legally and for free, read them without drowning, judge whether they can be trusted, and turn them into a short literature review. You don't need any background. If you can use a search engine and a word processor, you have everything you need to start.

What you'll be able to do when you finish

• Tell a primary research article from a review, a preprint, a conference paper, and a thesis, and know what peer review does and doesn't promise.
• Build a search strategy from your question, and run it well in Google Scholar, PubMed, and arXiv.
• Get the full text of almost any paper without paying and without breaking the law.
• Follow a good paper's citations backward and forward to find the rest of the field.
• Read a paper in three passes, and read its figures, error bars, and tables correctly.
• Spot weak evidence: no control group, tiny samples, retracted papers, predatory journals.
• Keep everything organized in Zotero, take notes in your own words, and write a literature review that synthesizes instead of summarizing.

How each chapter is built

Every chapter opens with what it covers and why it matters. Then come the ideas in plain words, real examples you can copy (search strings, emails, notes, paragraphs), and the mistakes to avoid. Search strings appear in gray boxes like this one, and you can paste them straight into the database the chapter is about.

"school start time*" AND sleep[mh]

Each chapter ends with a short Check yourself list. If you can answer those without looking back, move on. The last chapter follows one student topic from the very first search to a finished literature review, so you can see every step done for real.

What you need

A computer with a web browser, a free Zotero account (Chapter 12 walks you through it), and a public library card if you can get one. Your school library may also have database access you don't know about yet. Ask the librarian. They are some of the best search experts you will ever meet.

A suggested pace

Most students do this alongside choosing a research question, which the sibling guide "From Idea to Research Question" covers. The two guides feed each other: reading sharpens your question, and a sharper question makes searching easier. Your mentor can help you skip ahead to whatever your project needs first.

Before you search, you need to know what you're looking for. This chapter explains the main kinds of scholarly writing, what peer review is, how journals and DOIs work, and the parts every research paper has, so you can tell at a glance what a document is and where the useful parts are.

The kinds of scholarly writing

Not everything that looks like a paper is the same kind of thing. The type tells you how much to trust it and how to use it.

A primary source reports original work. A secondary source describes or analyzes other people's work. Review articles are secondary sources. News stories about a study are secondary too, and they often get details wrong, so always go back to the paper itself.

Peer review

Peer review is the check a paper gets before a journal publishes it. The journal's editor sends the manuscript to two or three independent experts, called referees or reviewers. They read it critically and recommend accepting it, asking for revisions, or rejecting it. Most papers go through at least one round of revisions. The whole process often takes months.

Journals run review in different ways. In single-anonymous review, the reviewers know who the authors are but the authors don't know the reviewers. In double-anonymous review, neither side knows. In open review, names or reports are made public.

Peer review catches many problems, but it is not a guarantee. Reviewers usually don't see the raw data or rerun the analysis. They can miss mistakes, and they can't detect well-hidden fraud. Read "peer reviewed" as "checked by experts once." Chapter 10 shows how to judge a paper yourself.

Journals, volumes, and DOIs

A journal publishes papers in a field, grouped into volumes (usually one per year) and issues. A citation like Pediatrics, 134(3), 642–649 means volume 134, issue 3, pages 642 to 649. Many journals now publish online-only articles with an article number like e2021054068 in place of page numbers.

Many papers appear online weeks or months before they are assigned to an issue. That's why the year in a database sometimes differs from the year on the printed issue. Use the year of the volume when you cite, and don't panic when two databases disagree by one year.

A DOI (Digital Object Identifier) is a permanent ID for a paper. It always starts with 10., then a prefix, a slash, and a suffix, like 10.1016/j.smrv.2015.06.002. Put https://doi.org/ in front of any DOI and it takes you to the paper's official page, even if the publisher moves its website. DOIs are the most reliable way to identify a paper, and Zotero can build a full citation from a DOI alone (Chapter 12).

Other IDs you'll see: a PMID is a paper's number in PubMed, a PMCID (starting with PMC) is its number in PubMed Central's free full-text archive, and an arXiv ID like 2501.01234 identifies an arXiv preprint.

The anatomy of a research paper

Most primary research articles follow the same structure, often called IMRaD: Introduction, Methods, Results, and Discussion. Once you know what each part is for, you can jump straight to the part that answers your question.

Here is what to expect from each part.

• Title. Usually states the main finding or the question. Titles with a colon often put the topic first and the finding second.
• Authors and affiliations. Where the authors work. One author is marked as the corresponding author, usually with an envelope icon or an asterisk, with an email address. That's who you contact with questions (Chapter 2). In many fields the first author did most of the work and the last author led the lab.
• Abstract. A summary of the question, methods, main results with numbers, and conclusion. Some journals use a structured abstract with labeled parts like Objective, Design, Participants, Results, and Conclusions.
• Introduction. Starts broad, narrows to what is already known, names a gap (something not yet known), and ends with the paper's question or hypothesis. The last paragraph of the introduction is often the most useful paragraph in the whole paper.
• Methods. Who or what was studied, how many, what was measured and how, and which statistics were used. Dense, but it's where you find out whether to trust the results.
• Results. The findings, mostly in figures and tables, with statistics in the text. Results report what happened without much interpretation.
• Discussion. What the results mean, how they compare with earlier studies, the limitations (weaknesses the authors admit), and what should be studied next. Limitations and future-work paragraphs are where student project ideas hide.
• Declarations. Funding sources, conflicts of interest, author contributions, and where the data and code are. Chapter 10 explains why these matter.
• References. The full list of works cited. Chapter 7 shows how to mine it.
• Supplementary material. Extra figures, full data tables, questionnaires, and code that didn't fit. It's usually a link near the end or on the article's web page. For the details of how something was measured, look here.

Some fields vary the pattern. Computer science papers often have "Related work," "Approach," and "Experiments" sections. Math papers are organized around definitions, theorems, and proofs. Review articles are organized by theme. The jobs are the same, only the headings change.