How to Use a Library for Research: Keywords, Databases & Sources

How to use a library for research becomes much easier when you stop treating the search box as a place to type a whole assignment question and start treating research as a sequence of small decisions. A strong library research workflow moves from a clear question to useful keywords, from keywords to catalogues and databases, from search results to credible sources, and from sources to notes, comparisons, citations and new questions. This guide shows students, parents and independent learners how to research a topic with library databases, Boolean search, subject headings, source evaluation and a practical evidence trail.

Library research is different from ordinary web searching because catalogues and databases are organised around records, subjects, authors, publications, dates and controlled vocabularies. That structure can feel unfamiliar at first, but it is exactly what makes libraries powerful. Instead of accepting whatever happens to rank first, you can decide what kind of evidence you need, search the right collection, narrow by date or format, compare several sources and trace a claim back to a book, article, report, newspaper or archival record.

In Singapore, the National Library Board’s current OneSearch and eResources are useful examples of this research environment. OneSearch brings together books and a much wider discovery layer, while NLB eResources provides access routes to subscribed databases, eJournals, eBooks, eNewspapers and other digital collections according to the resource’s access conditions. The exact tools will differ from one library to another, but the research habits in this article travel well: define, search, evaluate, connect, record and verify.

The 60-second answer: how do you research a topic in a library?

Turn the assignment or curiosity into one clear research question.
Break that question into two to four main concepts.
Write synonyms, related terms, broader terms and narrower terms for each concept.
Use the library catalogue for books and broad subject orientation.
Use library databases for journal articles, reports, newspapers and specialised material.
Search with combinations of keywords rather than one long sentence.
Open promising records and inspect subject headings, abstracts and references.
Evaluate who created the source, why, when, for whom and on what evidence.
Take notes in your own words and record citation details immediately.
Follow references and subject terms to discover the next layer of stronger sources.

This sequence matters because research is not one search. It is an iterative loop. Your first search teaches you the vocabulary of the topic. That vocabulary improves the second search. The second search reveals experts, institutions, dates and debates. Those clues improve the third search. Good researchers become faster not because they know a secret search engine, but because each result helps them formulate a better next query.

Start with a research question, not a pile of keywords

The easiest way to make a library search difficult is to start without deciding what you are trying to answer. “Renewable energy” is a topic. “How has rooftop solar changed electricity demand in dense tropical cities?” is a research question. “Shakespeare” is a topic. “How does Shakespeare use mistaken identity to create comedy in Twelfth Night?” is a research question. A question gives your reading a direction and gives every source a test: does this help answer what I am actually asking?

A useful research question is neither so broad that it could fill an encyclopedia nor so narrow that almost nothing has been written about it. If your question is “What is education?”, the library will return a universe. If your question is “What was the exact influence of one unnamed teacher on one unrecorded lesson in 1912?”, the evidence may not exist. Research begins by finding the workable middle.

Use four limiters to make a broad topic researchable

Time: current, historical, before or after a change, a specific decade.
Place: Singapore, Southeast Asia, a neighbourhood, a country, an ecosystem.
Population or object: Primary students, commuters, mangroves, businesses, novels, schools.
Aspect: cost, behaviour, achievement, design, policy, language, risk, culture, technology.

You do not always need all four, but asking about them helps. “Social media and teenagers” can become “How does short-form social video affect study concentration among secondary-school students?” The second question gives you concepts that can actually be searched: social video, adolescents, attention, study, concentration.

Turn the question into search concepts

Library databases are usually better when you search the important nouns and ideas rather than the whole natural-language question. Remove instruction words such as “analyse”, “discuss”, “compare” and “evaluate”. Keep the concepts that identify the subject of the research.

Suppose your question is: “How does access to public libraries affect reading habits among teenagers in cities?” The core concepts might be public libraries, teenagers, reading habits and perhaps urban. You can then build alternatives for each concept.

public libraries: library access, public library services, community libraries
teenagers: adolescents, youth, secondary-school students
reading habits: reading behaviour, reading frequency, recreational reading, reading engagement
urban: city, metropolitan, neighbourhood

This is where research begins to feel less like guessing. You now have a small vocabulary system rather than one fragile phrase.

Why synonyms matter more in databases than many beginners expect

General web search engines often infer related meanings, spelling variations and common synonyms automatically. Library databases may be more literal because they are matching fields such as title, abstract, subject or descriptor. If the author used “adolescents” and you searched only “teenagers”, you may miss a useful article. If one database uses “higher education” while another uses “tertiary education”, your search needs to be flexible enough to bridge the language.

Build synonym groups. Use OR inside each group and AND between different concepts. For example: (teenager OR adolescent OR youth) AND (reading habit OR reading engagement) AND library. The exact syntax varies between platforms, but the conceptual logic is stable.

Boolean search without the headache

AND narrows by requiring different concepts. OR broadens by accepting alternatives. NOT excludes, but use it carefully because you can remove useful records by accident. Parentheses group related alternatives. Quotation marks often keep an exact phrase together. Truncation symbols such as an asterisk may retrieve word variants in some databases, but the symbol and behaviour are platform-specific, so check the help guide before assuming every database works the same way.

A simple search string might look like: (“public library” OR “community library”) AND (teenager OR adolescent OR youth) AND (“reading habits” OR “reading engagement”). That is not a magic formula. It is simply a transparent statement of what must be present and what can vary.

When your search returns too many results

Add another concept with AND.
Search a key phrase in quotation marks.
Limit to title or subject fields when appropriate.
Filter by publication date when currency matters.
Filter by resource type such as peer-reviewed article, report or book.
Add a place, population or method.
Use a more specific subject heading discovered in a good record.

When your search returns too few results

Remove one restrictive concept.
Add synonyms with OR.
Try a broader term.
Remove quotation marks from phrases that may be expressed in several ways.
Search another database with better subject coverage.
Look at the vocabulary used in one relevant article and borrow its keywords.
Ask a librarian whether the database uses a controlled subject term you have not considered.

The catalogue and the database do different jobs

A library catalogue tells you what the library owns, licenses or can help you locate. It is especially good for books, eBooks, audiovisual material, journals as whole publications and sometimes local or archival collections. A research database usually indexes individual articles, chapters, conference papers, reports or specialised records. Searching the catalogue for “journal article about reading motivation” may be less efficient than searching an education or multidisciplinary database. Searching a database for “a beginner book explaining photosynthesis” may be less efficient than starting in the catalogue.

Think of the catalogue as the map of collections and databases as specialised microscopes. Use the tool that matches the material you want.

Start broad with background sources

Before chasing specialised journal articles, get enough background knowledge to understand the vocabulary of the field. Reference books, encyclopedias, introductory textbooks, handbooks and good overview chapters can save enormous time. They tell you what the topic is called, which names matter, what major subtopics exist and which debates are settled or contested.

This is particularly valuable for students entering unfamiliar subjects. If you immediately search “quantum decoherence applications” without knowing the difference between decoherence, entanglement and measurement, you will spend energy interpreting search results rather than researching. Ten minutes with a reliable overview can give you the words needed for the next hour.

Use subject headings as a shortcut into expert vocabulary

One of the best library skills is learning to notice subject headings, descriptors or controlled vocabulary. These are standard labels applied to records so that materials on the same concept can be grouped even when authors use different language. A title might say “young people”, the abstract might say “teens”, and the database descriptor might say “adolescents”. Once you discover the descriptor, you can search with the database’s own organised language.

A good routine is: find one highly relevant record, open the full record, inspect the subject or descriptor fields, click the most useful term, then browse the records gathered under it. This can turn one lucky discovery into a systematic search.

Read the result screen like a researcher

A result list is not merely a row of titles to click. It is a compressed map of the field. Scan authors, publication names, dates, document types and recurring words. If the same journal, organisation, theory or phrase appears repeatedly, make a note. Those repetitions tell you how the subject is organised and which vocabulary is likely to produce better searches.

Also notice what is missing. If every relevant result is ten years old, the topic may be mature, the database may not cover the newest material, or your current keywords may reflect older terminology. If every result comes from one discipline, ask whether another field studies the same problem under a different name. Search results are evidence about the search itself.

Read the abstract before you download everything

An abstract is a compact preview of a research article. It often tells you the question, population, method, main finding and implications. Use it as a screening tool. Ask: Is this directly about my question? Is the population relevant? Is the date suitable? Is this the type of evidence I need?

Students often download twenty PDFs because the titles look related. That creates a new problem: document overload. Read the title, abstract, keywords and publication information first. Keep only the sources that have a clear job in your project.

Give every source a job

A research project becomes manageable when each source plays a role. One source may define a concept. Another may provide historical background. A third may supply recent data. A fourth may challenge the dominant explanation. A fifth may provide a Singapore case. You are building an evidence architecture, not collecting trophies.

Definition source: What does the key term mean?
Background source: What context does the reader need?
Evidence source: What data or documented observation supports the claim?
Method source: How was the phenomenon studied?
Contrast source: What other interpretation exists?
Local source: What changes in the Singapore or regional context?
Primary source: What did a participant, institution, law, speech, dataset or historical document actually say or record?

Think in an evidence hierarchy, but do not make it rigid

Different questions need different kinds of evidence. A systematic review may be excellent for asking what a body of medical studies collectively suggests, but it cannot replace an original speech when your question is what a historical leader actually said. An official dataset may be ideal for unemployment figures, while an interview may be better for understanding how unemployment feels. A controlled experiment can isolate a causal mechanism, while an ethnography can reveal context the experiment deliberately removes.

Instead of asking “Which source type is always strongest?”, ask “Which source type is strongest for this claim?” Then triangulate when the claim is important. A policy question may require the policy document, outcome statistics, scholarly evaluation and stakeholder experience. A literary question may require the primary text, criticism and historical context. Good research matches evidence design to question design.

This is especially important for students who have learned a simple ladder where journals are always better than books and books are always better than websites. Quality does not live in the container. It lives in authorship, evidence, method, editorial process, relevance and fit. A current official website can be the definitive source for an application deadline. A peer-reviewed article from twenty years ago cannot tell you that deadline.

Books are slower than search results—and that is often useful

A strong book gives sustained structure. It can define a field, show relationships between ideas and build an argument across chapters. That is different from reading scattered fragments online. For school research, books are particularly useful for orientation, historical context, major theories and vocabulary. They may not be the best source for today’s statistics or a fast-moving software tool, but they can explain the frame in which newer information belongs.

Use the table of contents to decide which chapters matter. Use the index to locate your key term. Use the bibliography to identify foundational authors. You do not need to read every page to use a research book intelligently.

Journal articles are useful when you need focused evidence

Scholarly articles often answer narrower questions than books. They may report an experiment, survey, case study, model, review or theoretical argument. Their strength is precision. Their difficulty is density. Read them strategically: title, abstract, introduction, headings, figures or tables, discussion, conclusion, then return to the methods and details that matter for your purpose.

Do not confuse “peer reviewed” with “automatically correct”. Peer review is one quality-control process, not a guarantee that every conclusion is final. Compare findings, look at sample size, method, limitations and whether later research supports or challenges the claim.

Primary and secondary sources: ask what kind of evidence you need

A primary source is evidence created by participants, institutions or observers close to the event or phenomenon: speeches, letters, laws, raw datasets, interview transcripts, historical newspapers, photographs, maps, original research data, artworks or official records. A secondary source interprets, analyses or synthesises primary material. Both are valuable.

For a history project, a newspaper from the period can show contemporary language while a later historian can provide context and interpretation. For science, an original research paper can provide methods and data while a review article can show how the field fits together. Good research often uses both.

Use newspapers for the right reason

Newspapers are excellent for tracing public discussion, events, statements, chronology and how issues were framed at the time. They are less suitable when you need a settled scientific conclusion. Match the source to the claim. If the claim is “how did people talk about this event in 1985?”, a newspaper may be primary evidence. If the claim is “what is the current medical consensus?”, look for authoritative medical or scientific evidence instead.

Singapore researchers can use NLB discovery tools and appropriate digital newspaper resources where access permits. Historical newspapers can be especially powerful when combined with later scholarship because the contrast between contemporaneous reporting and retrospective analysis teaches you how historical interpretation works.

Evaluate a source by asking five better questions

Authority: Who created it and what expertise or institutional responsibility do they have?
Evidence: What data, documents, observations or references support the claims?
Purpose: Is the source explaining, persuading, selling, entertaining, reporting or advocating?
Currency: Does the age of the source matter for this question?
Fit: Even if the source is reliable, is it actually relevant to the population, place, level and claim in your project?

Source evaluation is not a ritual where you label a website “good” or “bad”. It is claim-specific. An old engineering manual may be excellent evidence of historical practice but poor guidance for current safety standards. A company website may be authoritative about its own product specifications but not neutral about whether that product is best. A government page may be primary evidence of policy, while an independent study may be needed to evaluate outcomes.

Authority is not the same as popularity

Search ranking, followers, views and shares are not evidence of expertise. Libraries teach a useful discipline: identify the author, publisher, institution and publication context. Ask what process placed the information there. Was it edited? Peer reviewed? Officially issued? Independently reported? Self-published? Automatically generated?

This does not mean self-published material is always useless. A personal testimony may be exactly what you need for lived experience. It simply means you must understand what kind of evidence it represents.

Use lateral reading when a source is unfamiliar

If you encounter an unfamiliar organisation, open new searches about the organisation rather than trusting only its own About page. Look for independent descriptions, funding information, citations by established institutions, corrections, criticisms and the wider reputation of the author or publisher. This habit is called lateral reading: you leave the page to learn what the web says about the source.

Library research and web verification work well together here. The library provides structured sources; the open web helps you inspect the context around publishers, organisations and current claims.

Do not confuse full text with best evidence

Databases sometimes tempt users toward whatever has an instant PDF button. But the most useful article may require a different access route, a linked journal page, another database or a request. “Available in one click” is a convenience criterion, not a quality criterion. Search for relevance first, then solve access.

If the library does not provide immediate access, check whether another edition, repository, open-access version, print holding or borrowing route exists. A librarian can often help you navigate these options.

Citation chaining: let one good source lead to ten more

When you find a strong article or book, look backward at its references. Which sources does it cite repeatedly? Which foundational studies or books appear important? Search those titles or authors. Then look forward where your tools allow: who later cited this source? That reveals how the conversation developed.

Citation chaining is powerful because scholars have already done some of the discovery work. You are following an intellectual trail rather than starting every search from zero.

Use publication date intelligently

“Newest” does not always mean “best”. For fast-changing topics such as AI tools, cybersecurity or medical guidance, recent information matters strongly. For philosophy, literature, mathematics or historical theory, older works may remain foundational. Ask why the date matters.

A good research set often contains both foundational and current material: an older source that established the idea and newer sources showing how it has been tested, revised or applied.

Keep a search log so you do not research in circles

A search log can be as simple as a note with four columns: database, search string, useful filters and what happened. Record searches that worked and searches that failed. If you tried “teenagers AND reading AND library” and the results were too broad, write that down before refining it. If adding “Singapore” removed almost everything, record that too. The log prevents you from repeating the same dead-end search two days later.

The log also makes your thinking visible. You can see how the question evolved. Perhaps you began with “screen time and learning”, discovered that researchers distinguish passive viewing from interactive use, and then changed the question to compare specific forms of screen activity. That change is not a mistake. It is evidence that the research process taught you something.

For larger projects, save the exact database name and search date because database contents and interfaces change. If your teacher asks how you found the evidence, you can explain the route rather than saying “I searched around”. That is a small but important move toward reproducible research.

Treat zero results as information, not failure

A zero-result search can mean several different things. Your phrase may be too exact. Your terminology may be unusual. The database may not cover the discipline. The topic may genuinely be under-researched. The date or place filter may be too narrow. Or the idea may be discussed under a different name. Each explanation suggests a different repair.

First, remove unnecessary quotation marks. Second, replace long phrases with key concepts. Third, add synonyms. Fourth, search a broader database. Fifth, find one background source and steal its vocabulary. If you still find almost nothing, ask whether the absence itself is meaningful. A research gap can be interesting, but only after you have ruled out a vocabulary problem.

How to research a Singapore topic without getting trapped in global generalities

Add local vocabulary early. Search Singapore, the relevant ministry or agency, neighbourhood names, school levels, local programme names, policy titles, local datasets and Singapore-specific subject headings. Then combine those with the broader academic concept. For example, instead of only searching “reading motivation”, try “reading motivation Singapore students”, then inspect which terms local researchers use.

Local history may require a different route from local science. For historical topics, maps, newspapers, oral histories, archival records and photographs can matter. For education, policy pages, curriculum documents, assessment frameworks and educational research may matter. For urban topics, planning documents, statistics, geospatial resources and historical photographs may matter. The library helps you move between these evidence types.

A research example: public libraries and teenage reading

Imagine a student asks: “Do public libraries increase reading among teenagers?” Start with background: what counts as reading engagement, library use and adolescent reading? Search the catalogue for books on reading habits and library services. Search education or information-science databases for studies. Add synonyms: adolescent, youth, teen, recreational reading, reading frequency, public library use. Add place terms if the assignment needs Singapore or another country.

Then evaluate the evidence. A survey showing that frequent library users read more does not automatically prove the library caused the reading. Perhaps people who already love reading visit libraries more often. Look for longitudinal studies, interventions or research that considers confounding factors. This is how a library project becomes reasoning rather than summary.

A research example: technology and learning

Suppose the question is: “Does generative AI improve student learning?” Do not search only “AI good for students”. Break it down: generative AI, students, learning outcomes, feedback, writing, mathematics, critical thinking. Search educational databases and recent reviews. Separate studies of perceived usefulness from studies measuring actual performance. Separate short-term task completion from long-term learning.

Then compare populations. University students using AI for programming may tell you little about Primary pupils learning vocabulary. A credible study can still be irrelevant to your exact claim. Relevance is part of source quality.

A research example: a literature essay

For a literature essay, the primary source is the text itself. Library research provides criticism, historical context, author scholarship, editions, dictionaries of literary terms and related works. Search the title, author and the technique or theme you are analysing. Add terms such as symbolism, narrative voice, irony, identity, setting or adaptation.

Do not let criticism replace your reading. Use scholarship to sharpen interpretation, not to outsource it. Quote the literary text for the evidence you can see directly and use secondary sources to show how other readers frame the question.

A research example: science inquiry

For science, begin with a textbook or reliable overview to stabilise definitions. Then move to review articles and original studies if the level requires it. Search by organism, variable, process, method and environment. Read figures carefully. Ask whether the experiment measured what the headline claims.

A student researching microplastics might need environmental sampling studies, laboratory toxicity experiments and policy reports. Those are different evidence streams. The research question should determine which matters most.

Research notes should preserve the source trail

Never take an important note without recording where it came from. Write author, title, publication, year, page number or stable link as you work. It is much easier to remove extra citation information later than to reconstruct a missing source at midnight.

Separate three things in your notes: direct quotation, paraphrase and your own idea. Use quotation marks around exact wording. Paraphrase only after you understand the point. Label your own analysis clearly so it does not accidentally become attributed to the source.

A practical note template

Question or sub-question this source helps answer
Source details
Main claim
Evidence or example
Useful quotation with page number if needed
Limitations or cautions
Connection to another source
My interpretation
Next search term

Why “next search term” belongs in your notes

Every good source teaches you the vocabulary of the field. An article may introduce a theory name. A book may mention a government programme. A database record may use a subject heading you had not considered. Recording the next search term turns reading into discovery. It is the research equivalent of leaving breadcrumbs for your future self.

Use a source matrix when the project grows

For a longer project, create a simple table with sources down one side and themes across the top. Mark which source supports each theme, contradicts another, provides data, uses a particular method or addresses a different population. A matrix lets you compare rather than summarise one source at a time.

This prevents the “book report” structure where paragraph one is Source A, paragraph two is Source B and paragraph three is Source C. Strong research writing is organised by ideas. Multiple sources meet inside each idea.

Protect citation integrity from the first note

Citation problems often begin long before the bibliography is written. They begin when a student copies a sentence into notes without quotation marks, saves a screenshot without its source, or writes down a statistic without the year and denominator. By the time the essay is drafted, the memory of what came from where has blurred. The repair is simple: preserve provenance at capture time.

If you quote, record the exact page. If you paraphrase, keep the source attached. If you use a table, note the table title and underlying dataset. If an online page changes, record the organisation and access route required by your citation style. Citation is not decoration added after writing. It is part of the evidence system that lets another reader trace your reasoning.

This discipline also protects you when using AI. If an AI assistant suggests a claim, treat it as a lead until you can connect it to a real source you have inspected. Do not create a citation merely because a generated answer supplied a plausible author and title. Search the library or publisher record, verify that the source exists, and verify that it actually supports the claim.

How to search faster without becoming shallow

Speed comes from reducing useless actions. Save strong searches. Reuse synonym groups. Learn the subject vocabulary. Scan abstracts before downloading. Search within documents. Use indexes. Keep citation details. Stop reading a source when it clearly does not fit. Ask for help when the problem is the search strategy rather than the topic.

Do not measure productivity by how many tabs are open. Measure it by whether you can now answer the question more precisely.

The 20-minute research warm-up

Minutes 0–3: write the question and the assignment requirements.
Minutes 3–6: identify three main concepts and synonyms.
Minutes 6–10: search the catalogue for one background book or handbook.
Minutes 10–14: search one suitable database using a simple Boolean string.
Minutes 14–17: open the best record and inspect abstract, subjects and references.
Minutes 17–20: write three new search terms and one revised version of the research question.

After twenty minutes, you should not have “finished the research”. You should have a better map. That is a much more useful milestone.

The 90-minute library research session

First 15 minutes: orientation. Clarify question, scope and vocabulary.
Next 20 minutes: catalogue and background sources.
Next 25 minutes: database searching and screening.
Next 15 minutes: read the strongest two sources deeply.
Next 10 minutes: compare claims and note gaps.
Final 5 minutes: save citations, record next searches and decide what evidence is still missing.

This structure prevents one of the most common failures in student research: spending the whole session searching and leaving no time to understand what was found.

When to stop searching

Research can expand forever. Stop when new sources are mostly repeating what you already know, your main claims have appropriate evidence, major counterarguments are represented and the assignment’s scope is satisfied. If every new source introduces a completely new dimension, your question may still be too broad.

Stopping is a research skill. The goal is not to know everything. It is to build a defensible answer to a defined question.

How AI can support library research without replacing source work

AI is useful for brainstorming synonyms, turning a broad interest into possible questions, explaining unfamiliar terminology, generating comparison prompts and helping you organise notes you already possess. It is much less reliable as an invisible substitute for source discovery. Generated citations can be wrong, incomplete or mismatched, and generated summaries can flatten disagreement.

A strong workflow is: use AI to improve the question, use library systems to find identifiable sources, read those sources, then use AI again to challenge your explanation or suggest gaps. Keep the evidence chain visible. If a factual claim matters, know which source supports it.

Why libraries matter more in an age of abundant information

Scarcity used to be the problem: not enough access to information. Abundance creates a different problem: too much information with uneven quality, uncertain provenance and competing incentives. Libraries respond with structure. Catalogues identify items. Databases provide fields and filters. Subject headings organise concepts. Librarians help users translate questions into search strategies. Citation systems preserve traceability.

The modern library is therefore not merely a warehouse of answers. It is an environment for learning how to move through evidence.

Connect this research workflow to the wider eduKate learning graph

For the physical navigation layer, begin with How to use a Library | Find Books Fast with the Catalogue, Call Numbers and Shelves. For the larger institutional picture, continue to How Public Libraries Work. For source judgement, see The Gold Standard Of Information Literacy. For a broader research skill framework, connect with How to be Good at Research.

The purpose of these links is not to send the reader in circles. Each article owns a different layer. This one is the search-and-evidence engine: how a question becomes a defensible set of sources.

Frequently asked questions

What is the difference between a library catalogue and a database?

A catalogue primarily describes items the library holds or provides access to, such as books, eBooks, journals and media. A database often indexes individual articles, reports, chapters or specialised records. Use the catalogue for collection discovery and databases for focused research within subject literature.

Should I use Google Scholar instead of library databases?

Use both when appropriate. Google Scholar can be excellent for broad scholarly discovery and citation connections. Library databases may offer stronger subject indexing, controlled vocabulary, filters and access to subscribed full text. The best tool depends on the topic and source type.

What does peer reviewed mean?

It usually means a scholarly manuscript was evaluated by other specialists before publication. The exact process varies by journal. Peer review improves quality control but does not make a paper infallible. Evaluate methods, evidence, limitations and later research as well.

How many sources do I need?

Enough to support the claims required by the assignment. A short school project may need a small set of strong, varied sources. A major research paper may require many more. Quality, relevance and coverage matter more than an arbitrary pile.

What should I do if every search gives irrelevant results?

Return to vocabulary. Find one good background source, inspect its subject headings and terminology, then rebuild the search. If that still fails, ask a librarian to help identify the right database or controlled term.

Can I trust an official source?

Official sources are often authoritative for what an institution states, records or regulates. That does not mean they answer every evaluative question neutrally. Use the official source for the claim it is qualified to make, then add independent evidence where needed.

What is the best first database for a beginner?

There is no universal best database. A multidisciplinary database can be a reasonable start, but subject databases become more powerful once you know the field. Library research guides and librarians are designed to help you choose.

How do I know whether a source is too old?

Ask whether the claim changes over time. Current software, medical guidance, policy and statistics may age quickly. Historical interpretation, foundational theory and literary criticism may remain relevant for decades. Date is a relevance question, not a simple expiry label.

What if the article I need is not available in full text?

Search the title in the catalogue and other databases, look for an open-access version or repository copy, and ask the library about document-delivery or borrowing options. Do not automatically replace a perfect source with a weaker one simply because the weaker PDF opens faster.

What is the most important library research habit?

Let every useful source improve your next search. Research accelerates when reading and searching feed each other.

The research checklist

My question is specific enough to investigate.
I identified the main concepts and useful synonyms.
I used the catalogue for books and orientation.
I used an appropriate database for articles or specialised sources.
I tried subject headings or descriptors.
I screened titles and abstracts before downloading.
I know what job each source performs.
I evaluated authority, evidence, purpose, currency and fit.
I recorded citation details and page numbers while working.
I compared sources rather than summarising them one by one.
I followed references and new vocabulary to stronger material.
I know what evidence is still missing.

Use a library as a research system. Start with a question, learn the language of the topic, search the right collection, evaluate the evidence, preserve the source trail and let each answer produce a better next question. That is how a library turns information into knowledge.

How to use a Library | Research a Topic with Keywords, Databases and Credible Sources