Reference Managers and Citation Styles: EndNote, Zotero, Mendeley, Vancouver and Harvard

Forty-Seven Minutes on a Reference List

Dr. Hammad Ali came into the Research Room with his laptop open and his face the colour of a failed blood culture.

“Sir, my supervisor sent my introduction back. He did not even read the introduction. He read the references and stopped.”

He turned the screen towards me. Eleven references. Four had numbers, three had author names and years in brackets, two were website links, one was a journal name with no article, and the last one simply said “Ahmed et al., Lancet”. Reference 3 in the text pointed to a paper that appeared as number 7 in the list, because he had added a new paragraph and renumbered everything by hand.

“How long did this take you?” I asked.

“Forty-seven minutes, sir. Just the references. And then I added one more paper and everything moved again.”

Dr. Junaid Rashid laughed from the corner. “Fifteen years ago I wrote my only case report the same way. I typed every reference myself. The journal returned it twice for formatting.”

“Then today you will both learn the thing that would have saved those hours,” I said. “Nobody should be typing references in this decade. A reference manager does it for you, and it does it correctly, every time, in whatever style the reader wants.”

Two Jobs, Two Different Things

Most trainees mix up two separate ideas, so I wrote them on the whiteboard.

• A reference manager is software. It stores the details of every paper you read (authors, title, journal, year, volume, pages, DOI, PMID), keeps the PDF with it, and inserts citations into your Word file.
• A citation style is a set of rules. It decides how a citation looks in your text and how the reference list is written at the end. Vancouver and Harvard are the two you will meet most.

“The software knows thousands of styles,” I said. “You choose the style once. The software applies it to every reference. If the journal wants a different style next month, you change one setting and the whole list rewrites itself in seconds.”

Hammad looked at the forty-seven minutes on his screen and said nothing, which for Hammad is a strong emotion.

The Three Tools: EndNote, Zotero and Mendeley

Dr. Hassan Raza had his pen ready. “Sir, just tell me which one is best.”

“The best one is the one you will actually use,” I said. “But here is the honest comparison.”

	EndNote	Zotero	Mendeley
Made by	Clarivate	Non-profit, open-source project	Elsevier
Cost	Paid licence (some universities provide one)	Free	Free
Works with Word	Yes, Cite While You Write	Yes, Zotero Word plugin	Yes, Mendeley Cite add-in
Captures papers from the browser	Yes	Yes, Zotero Connector	Yes, Mendeley Web Importer
Shared libraries for co-authors	Yes	Yes, free group libraries	Yes
Best for	Large theses and long-term libraries, if your institution has a licence	Most trainees: free, reliable, easy to learn	Those already using Elsevier tools

EndNote is the oldest and most powerful. It handles thousands of references without complaint and many senior researchers use it. Its weakness for a Pakistani trainee is simple: it costs money. If your university or hospital library provides a licence, use it.

Zotero is free and open source. It comes from a non-profit project, not from a publisher. The free account syncs your library across computers and gives you a limited amount of cloud space for PDFs, which is enough for a synopsis and a dissertation if you are sensible. For most of my trainees, Zotero is the one I recommend.

Mendeley is free and owned by Elsevier. One warning: the old Mendeley Desktop has been retired. If a senior tells you to “download Mendeley Desktop”, use the current Mendeley Reference Manager with the Mendeley Cite add-in for Word instead.

Dr. Zunaira Malik raised her hand quietly. “Sir, can I start in one and move later?”

“Yes. All three can export and import libraries, usually as an RIS file. But do not keep your references in two tools at the same time. You will end up with two versions of the truth, and one of them will be wrong.”

Zunaira Sets Up Zotero in Twenty Minutes

Zunaira chose Zotero, and the group watched her set it up on the projector. These are the steps, and they are almost the same in the other two tools.

• Install the desktop program and the browser connector. Download both from zotero.org. The Word plugin usually installs with the desktop program. If the Zotero tab does not appear in Word, restart Word once.
• Create a free account and switch on sync. If your laptop dies the week before submission, your library is still safe.
• Make one collection per project. Zunaira named hers OSMF_Areca_MPhil. Junaid later made HTN_Adherence_OPD.
• Add papers the clean way. On a PubMed page, click the connector button in the browser. Or use “Add Item by Identifier” and paste the PMID or DOI. The details come straight from the database instead of from your typing.
• Attach the PDF to each reference when you have the full text, so the paper and its details stay together.

“Notice what she did not do,” I said. “She did not copy references from Google Scholar, from another thesis, or from the reference list of another paper. Each reference came from the database record of the paper itself.”

Dr. Bushra Fatima nodded slowly. She had once copied a whole reference list from a paper she admired, and two of those references did not say what that paper claimed they said. She had not opened them. In this programme, the rule is simple: you cite only what you have read.

Check Every Imported Reference

Reference managers are only as good as the data they import. I asked Zunaira to open five of her new references and look closely. Within two minutes she found three problems:

• A title in capital letters throughout, because the journal website stored it that way.
• A paper marked “Epub ahead of print” that now had a volume and page numbers, which the record did not show.
• An author’s name entered as first name and surname swapped.

“Clean each reference once, when you add it,” I said. “Then it stays clean in every document you ever write. Check the authors, the title, the journal abbreviation, the year, the volume, the issue and the pages. If a field is empty, find it in PubMed.”

And one sentence for anyone tempted to ask a chatbot for references: we covered this in the AI session, and the rule has not changed. Every reference must come from a real record you have opened yourself.

Vancouver and Harvard: The Same Paper, Two Faces

Dr. Sumaira Talib and Hassan were sitting next to each other, which was useful, because their two universities ask for two different styles. Sumaira writes for the CPSP, which expects Vancouver. Hassan writes for UHS, which expects Harvard.

I put one real paper on the board, a guideline Sumaira will cite in her surgical site infection study, and wrote it in both styles.

Vancouver (numbered, used by most medical journals and set out in the ICMJE Recommendations and the National Library of Medicine’s Citing Medicine) (1, 2):

In the text: Surgical site infection remains one of the commonest healthcare-associated infections after surgery (1).

1. Berríos-Torres SI, Umscheid CA, Bratzler DW, Leas B, Stone EC, Kelz RR, et al. Centers for Disease Control and Prevention guideline for the prevention of surgical site infection, 2017. JAMA Surg. 2017;152(8):784-91.

Harvard (author and year, used by UHS and many universities):

In the text: Surgical site infection remains one of the commonest healthcare-associated infections after surgery (Berríos-Torres et al., 2017).

Berríos-Torres, S.I., Umscheid, C.A., Bratzler, D.W., Leas, B., Stone, E.C., Kelz, R.R. et al. (2017) ‘Centers for Disease Control and Prevention guideline for the prevention of surgical site infection, 2017’, JAMA Surgery, 152(8), pp. 784-791.

“This guideline is older than the recent references a CPSP synopsis expects,” I added. “I am using it only to show the format. Your actual list should be mostly recent.”

Then we compared the two side by side.

	Vancouver	Harvard
Citation in text	A number: (1), [1] or superscript, as the journal or institution asks	Author and year: (Ali, 2023) or (Ali and Khan, 2023)
Order of the reference list	In the order each source is first cited	Alphabetical by first author’s surname
Citing the same paper again	Use its original number	Repeat author and year
Many authors in the list	First six, then “et al.”	Depends on the Harvard version; follow your institution’s sample
Journal name	NLM abbreviation, e.g. JAMA Surg	Usually written in full, often in italics
Who asks for it here	CPSP synopsis and most medical journals	UHS synopsis and thesis

Hassan frowned. “Sir, which Harvard? I found five different versions online.”

“Good question, and the honest answer is that Harvard has no single official version,” I said. “Commas, full stops and italics differ from one guide to the next. So you follow the sample in the current UHS guidelines and whatever your department has accepted before. Choose the Harvard style in your reference manager that matches that sample, and then never touch the formatting by hand.”

What the CPSP and UHS Expect

These are the conventions our trainees work to. Confirm the current version with your supervisor and the latest guidelines before you submit, because formats change.

• CPSP synopsis: Vancouver style, about 10 to 12 references, recent papers from 2022 onwards.
• UHS synopsis: Harvard style, about 18 to 22 references, because the UHS format includes a literature review section.
• Similarity and AI reports: UHS expects the Turnitin similarity and AI reports below 15 percent; for the CPSP, below 19 percent. A clean, correctly formatted reference list does not lower similarity by itself, but references typed carelessly by hand are one more place for errors to creep in.

“Ten to twelve is not a target to fill,” I told Sumaira. “Every reference must support a sentence in your text. If you cannot point to the sentence, remove the reference.”

Citing While You Write

Junaid had been watching quietly. “Show me what happens in Word,” he said. “That is where I lost my weekends.”

I opened a blank document and typed one sentence from his introduction about hypertension in Pakistan. I placed the cursor at the end, opened the Zotero tab, clicked “Add/Edit Citation”, typed two letters of the author’s surname and pressed Enter. A number appeared. I clicked “Add/Edit Bibliography”, and a perfect Vancouver list appeared at the end. Then I moved the second paragraph above the first. Every number in the text and every entry in the list renumbered itself.

Junaid sat back. “Fifteen years,” he said. “Fifteen years I did that by hand.”

The rules that keep this working:

• Never type or edit a citation by hand inside a document that a reference manager is managing. The software will overwrite your change, or worse, get confused.
• Choose the style once, in Document Preferences, before you start: Vancouver for Sumaira, Harvard for Hassan.
• Keep the live version. When the document is final, save a copy, and only in that copy convert citations to plain text if your supervisor or the journal asks for it. Keep the original with live citations for future changes.
• Share the right way. If two people are writing together, use a shared group library so both of you insert citations from the same source.

That last point mattered more than I expected. Sumaira and Bushra had agreed, after months of friendly rivalry, to write a short review together on infection prevention in obstetric and surgical wards. Within ten minutes they had a shared Zotero group library. “If one of us fixes a reference,” Sumaira said, “it is fixed for both of us.” Bushra smiled for the first time that day.

Mistakes I See Every Month

• Mixing styles in one list. Hammad’s list had three styles. A reviewer sees this in five seconds and assumes the rest of the work is as careless.
• Numbers that do not match. The citation says (3) and the list says something else at number 3. This happens every time references are numbered by hand.
• Citing a website or Wikipedia for a medical fact that a guideline or a published study can support.
• Citing from another paper’s reference list without reading the original.
• Old references only. A CPSP synopsis with references from ten years ago tells the reviewer the literature was never searched properly.
• Missing details: no volume, no pages, the wrong year, or “et al.” after only two authors.
• Leaving references in the list that are never cited in the text, or the reverse.

Before you submit anything, run one final check. Count the citations in the text and the entries in the list. Open three references at random and confirm that each one says what your sentence claims it says. That five-minute check has saved more synopses than any software.

Hammad, One Hour Later

By the end of the session, Hammad had deleted his old list. He imported all eleven papers into Zotero by PMID, found that two of them were not about shoulder-tip pain at all, and removed them. He found two recent trials on low-pressure pneumoperitoneum that he had not read, read them properly, and added them. He chose Vancouver, inserted his citations, and generated the list.

“Six minutes, sir,” he said. “And I can add a paper without breaking anything.”

“Research is not rocket science,” I said. “It is a skill, and skills can be taught. Today you learned the one that saves you a weekend every month.”

Key Takeaways

• A reference manager stores your papers and builds citations; a citation style is the set of rules for how they look. Choose the tool once, choose the style per document.
• Zotero and Mendeley Reference Manager are free; EndNote is paid but powerful if your institution provides a licence. Use one tool only, and keep it synced.
• Import references from PubMed by PMID or DOI, attach the PDF, and clean each record once when you add it.
• Vancouver uses numbers in order of first citation (CPSP, most medical journals); Harvard uses author and year with an alphabetical list (UHS). Follow your institution’s sample for Harvard.
• Never type or edit citations by hand in a managed document, and keep a copy with live citations.
• Cite only papers you have opened and read, and check the text against the list before you submit.

References:

1. International Committee of Medical Journal Editors. Recommendations for the conduct, reporting, editing, and publication of scholarly work in medical journals [Internet]. ICMJE; updated January 2025. Available from: https://www.icmje.org/recommendations/
2. Patrias K, Wendling D, editors. Citing Medicine: the NLM style guide for authors, editors, and publishers [Internet]. 2nd ed. Bethesda (MD): National Library of Medicine (US); 2007-. Available from: https://www.ncbi.nlm.nih.gov/books/NBK7256/
3. Berríos-Torres SI, Umscheid CA, Bratzler DW, Leas B, Stone EC, Kelz RR, et al. Centers for Disease Control and Prevention guideline for the prevention of surgical site infection, 2017. JAMA Surg. 2017;152(8):784-91.