Practical write-ups on literature review, reading papers, writing, and putting AI to work in research.
A complete annotated bibliography example — three real entries written out in full, a breakdown of what each sentence in an annotation is doing, and the same entry set out in APA, MLA, and Chicago.
There is no single best paper search site. Here is what Google Scholar, PubMed, arXiv, Semantic Scholar, OpenAlex and the national indexes are each good and bad at, and how to combine two or three of them into a search you can defend.
A decision guide for EndNote versus Zotero — institutional licensing and support, what it costs to leave, Word and LaTeX workflows, and collaborating outside your institution — plus the job neither tool is built for.
The fastest routes to a BibTeX entry — one curl command against the DOI, the export links on arXiv and ACL Anthology, or a one-press copy from your own library — plus the six things that are usually wrong with whatever you get back.
Every reference answers the same five questions, and the style only decides the order and the punctuation. Worked examples for journal articles, books, chapters, web pages and theses in APA, MLA and IEEE, plus how in-text citations map to the list.
The two routes to a reference list in LaTeX — a .bib file with BibTeX, and biblatex with Biber — with working code, the cite commands, the natbib styles, and the exact compile order that makes the list appear. A third route, without a .bib file at all, comes at the end.
Keyword search finds papers that share your vocabulary. Literature mapping follows citations instead, which is how you find the work that matters but uses different words. Here is the method, and when to stop.
An outline decides the order of your claims, not the order of your papers. Here are the four ways to organize the body, what belongs in each section, a skeleton you can copy, and three tests that catch a broken outline before you write 5,000 words.
What it actually takes to connect a Zotero library to Claude or ChatGPT over MCP — the local API, the third-party servers, a worked config — and when a per-paper handoff is the better fit.
A practical comparison of Zotero and Mendeley for researchers — storage, plugins, LaTeX workflows, group libraries, and who owns your data — plus the one job neither of them does.
A repeatable way to run a literature review: fix the question first, collect in one place, take notes against the question, and only then write the synthesis.
Stop reading papers front to back in one sitting. Read them three times instead — once to discard, once to get the claim, once to check whether you could reproduce it.
Choose an AI research tool by the kind of service it is, not by the corpus size it advertises. The counts are measured differently, none is independently audited, and one vendor publishes the same figure at two precisions on the same day. Here are the three kinds and what each one misses.
How old is a paper when somebody cites it? We measured it across 19,379,997 papers and 413,582,341 citations. Arts and humanities are cited 13.3 years after publication and mathematics 12.5, energy at 5.0 years and engineering at 5.8. Here is the table for all 26 fields, and where the gap comes from.
A research gap is a question the existing studies fail to answer, for a reason you can name. Here is the four-reason framework from the systematic review literature, how to use a future-work section as a starting point, and how to trace citations to check your gap is still open.
An h-index is a count you can do by hand in about a minute from a citation-sorted list of papers. The catch is that the count belongs to the database you counted in, so the same researcher has several at once. Here is the definition, a worked example, where the numbers diverge, and what the index cannot tell you.
A step-by-step method for finding prior research, from choosing search terms to following citations in both directions. Includes what to do when a search returns nothing and when it returns too much.
Grey literature is research that never enters commercial publishing — theses, technical reports, working papers, government publications. Here is where the name comes from, the international definition, why it is worth the extra searching, where to actually look, and how to cite something that may quietly disappear.
The order to read a paper in, how to look up fewer words, and where machine translation helps and where it misleads. Replaces reading everything at the same depth with reading only what you need.
Installing Zotero, importing papers, organising them into collections, inserting citations in Word, and producing a reference list. Includes what to do when an import fails.
MeSH is the controlled vocabulary the US National Library of Medicine assigns to every paper it indexes. It collapses every way authors phrase the same condition into one heading, so synonyms stop costing you results. Here are its three parts, how to search with it, and what it cannot reach.
How to load paper PDFs into NotebookLM and ask questions about them. Covers how it shows its sources, the limits on how much you can add, and why it is not a tool for finding papers.
PRISMA is the guideline for reporting a systematic review: a 27-item checklist, a 12-item abstract checklist, and the flow diagram. Here is what each of the diagram's three tiers has to report, why records, reports and studies are counted separately, what changed in 2020, and how to pick an extension.
A six-field template for summarising one paper, and how to lay several papers side by side in a table. Produces notes you can turn straight into the literature review section.
Tools that search the literature before answering have largely closed off invented references. They have not closed off misreading a real paper, and the measured form of that misreading is overgeneralisation. Here is what to hand the model, the three places to check its output, and where the responsibility sits.
Three ways to use a Zotero library from Obsidian, from plugins to exporting BibTeX to not connecting them at all. Includes setup steps and the parts that break.