How to Conduct Legal Research: A Step-by-Step Beginner Guide

Finding a case feels like research. Confirming it is still good law is the step most attorneys skip, and the one where malpractice claims begin.

A significant share of solo practitioners lack access to Westlaw or LexisNexis, according to 2024 small-firm research. That number alone reframes the legal research problem. The real gap is structural, and the consequences land hardest on the attorneys who can least afford them.

Finding a case that matches your issue feels like progress. It looks like progress. But a case that was good law in 2019 may have been quietly overruled by 2023, and no keyword search will warn you. The retrieval step surfaces candidates; the verification step determines whether any of them are safe to cite. Attorneys who treat those two steps as one are, without realizing it, skipping the gate that protects them from citing bad law.

Solo attorney laptop showing a found case flagged as overruled bad law

The retrieval step surfaces candidates; the verification step determines whether any of them are safe to cite. That gap between "found it" and "confirmed it" is where most research errors occur.

Not because the attorney searched carelessly, but because the process they were taught stopped one step too early. LexisNexis's 2024 analysis of small-firm practice makes the structural disadvantage explicit: solo and small-firm attorneys operate under the same duty of competence as BigLaw partners, but without enterprise database access, mentorship pipelines, or peer-review layers to catch a bad cite before it reaches a filing. At a large firm, a senior associate or supervising partner is a backstop. At a solo practice, there is no backstop.

General AI tools seem to close the access gap. They don't. LexisNexis's 2024 guidance on generative AI in legal research identifies two distinct risk categories facing solo and small-firm attorneys:

The access problem: operating without enterprise database access, mentorship pipelines, or peer-review layers.
Fabricated or unverifiable citations: a risk separate from, and in addition to, the access problem, where a general chatbot may return a citation that looks precise, reads persuasively, and does not exist.

The access problem calls for a tool that grounds every answer in real, verifiable law rather than fluent approximation, one that treats verification as part of retrieval, not an afterthought the attorney handles alone after the search ends.

Key takeaways

A significant share of solo practitioners lack access to Westlaw or LexisNexis, meaning the legal research problem is structural, not just a skills gap.
Malpractice claims rarely start with a failure to find a source; they start with a failure to frame the issue correctly before searching anything.
Ten minutes of written issue-framing before your first query is a malpractice-prevention tool, not a time-management habit.
Free sources can close the access gap, but only if you sequence them correctly, starting broad and narrowing to binding primary authority.
Legal research fails most often at the sequence step: researchers who skip straight to case law miss the controlling statute sitting one layer above every case they find.
Retrieval and validation are two separate acts, a citation that was valid when you found it can become professionally dangerous the moment you file it without running a citator check.
94% of lawyers already use AI for research, but 83% don't trust it, because general-purpose tools can't self-correct the verification gap that creates real exposure.
OpenCase closes that gap by delivering accurate answers grounded in real law, so legal teams can research, draft, and review faster without trading speed for reliability.

What Legal Research Actually Covers Before You Search Anything

Solo lawyer tracing a chain of legal sources from facts to controlling case authority

What Legal Research Is - A Plain-Language Definition

The goal is a traceable line from your client's facts, through applicable statutes or regulations, to controlling case law in the right jurisdiction, ending at authority a court will accept.

LexisNexis's 2024 guidance for small firms puts it plainly: legal research requires verifying that cited authority is still good law, understanding how courts in the relevant jurisdiction have applied it, and confirming no subsequent statute or ruling has altered its effect. OpenCase surfaces this step directly: searching across 100+ legal databases, including integrated access to Cornell LII, the Federal Register, and daily PACER updates, means a practitioner can trace a case from its original citation through subsequent treatment in one connected workflow rather than toggling across disconnected platforms.

100+

legal databases searched in one pass

Primary vs. Secondary Sources - Why the Distinction Determines Your Research Order

Secondary sources explain and synthesize the law; they orient you in an unfamiliar area. Primary sources, statutes, regulations, and court opinions, are the law itself. OpenCase's statute analysis and file analysis capabilities are designed for that scenario: a practitioner can move from secondary orientation to primary-source verification within the same session, rather than treating them as separate research phases that each demand a different tool or database login.

Why Precision Beats Volume - Focus on Controlling Authority, Not Source Count

The attorneys who produce reliable work product fastest are not the ones who find the most sources. The ABA Journal's analysis of legal malpractice trends confirms that errors in case analysis, not just failure to locate authority, drive professional liability claims.

Searching across 100+ legal databases through OpenCase closes the gaps that narrow searches leave open. And because OpenCase integrates directly with Microsoft Word, Google Docs, Google Drive, Dropbox, and Outlook, the transition from verified research to drafted work product, whether a brief, a memo, or a client letter, does not require re-entering sources or switching contexts.

How to Build a Legal Research Plan That Keeps You on Track

The hidden cost isn't the time you spend searching; it's the write-downs that follow when those analytical gaps surface, a pattern Law Practice Magazine (American Bar Association) identifies as one of the quietest revenue leaks in solo practice.

Solo attorney's desk with a structured legal research plan notebook and AI platform on laptop

Pros and cons at a glance

Free federal and state court opinions → Pro: Includes citation history → Con: Coverage of older state appellate decisions is uneven.
Searchable case law database at no cost → Pro: Free access to searchable case law → Con: Citation-history features are not a substitute for a formal citator.
Citation history → Pro: Shows that a case has been cited → Con: Does not necessarily tell you whether the case was overruled.

The attorneys we work with at OpenCase face a version of this pressure every week: they need to build the research record for a memo without billing twenty hours to it, and they need the output to be defensible enough that a skeptical partner, or a bar complaint, can't pick it apart. That tension is exactly what a structured research plan is designed to resolve.

Write Your Legal Issue as One Sentence Before You Search Anything

One Legal frames issue-framing as the step that prevents scope creep. "Employment law in California" is a topic, while "whether a California employer's failure to provide written meal-period waivers creates liability under Labor Code section 512" is a research issue.

OpenCase's statute analysis and file analysis features are built around this same logic: the platform works from the specific legal question you've defined, surfacing the governing statute, regulation, or court rule that controls that situation, not a broad sweep of tangentially related material.

Lock Down Jurisdiction and Key Facts So Your Search Has a Fixed Address

This is exactly the problem OpenCase's Federal Register Integration and Daily PACER Integration address in practice, pulling jurisdiction-specific regulatory updates and docket activity so your fixed address stays current rather than drifting as the law moves.

Build a Search Vocabulary and Research Log Before Your First Query

One Legal recommends building your vocabulary first: legal terms of art, plain-language synonyms, and Boolean operators that tighten or broaden results deliberately rather than accidentally. Alongside that vocabulary, maintain a running research log from the moment you run your first query.

A simple table works: columns for source name, citation, jurisdiction, date accessed, and validation status. When those queries run inside OpenCase, which searches across 100+ legal databases in a single pass, including Cornell LII integration for primary statutory authority, the log populates with sourced, citable results rather than browser-history fragments. That matters when you need to prove to a skeptical partner that the research can be trusted: a timestamped log of 100+ database sources is a different conversation than "I searched Westlaw and some free sites."

OpenCase integrates with Microsoft Word, Google Docs, Outlook, Google Drive, and Dropbox, so the memo you build from your research plan moves directly into the workflow your firm already uses, without a reformat step that introduces its own error risk.

Once your plan is written, issue statement, jurisdiction, key facts, and search vocabulary locked, the next question is where to actually run those queries, especially if Westlaw and Lexis aren't in your budget. The answer starts with knowing which free and low-cost platforms are worth your time.

Where to Start Legal Research When You Don't Have Westlaw or Lexis

Subscription rates for Westlaw and LexisNexis can exceed $500 per month for a solo practitioner, according to industry research, and for attorneys running a one-person practice, that figure rarely survives a budget review.

Solo lawyer's free legal research flow from treatises to free databases to AI platform

Start With Secondary Sources Not Cases and Here Is Why That Is Faster

As most practitioners working without a subscription quickly discover, secondary sources such as treatises, practice guides, and legal encyclopedias supply the conceptual framework and direct citations to controlling primary authority. They are the fastest path to competent research. Skipping this step multiplies unproductive searches and converts free tools into expensive ones.

A treatise chapter or an American Jurisprudence entry on your issue will tell you the governing legal standard, the leading cases, and the statutory hook, often in four pages. Solo practitioners who work with OpenCase can run that same orientation query across 100+ legal databases in a single search, reaching practice guides, encyclopedias, and primary sources simultaneously without toggling between tabs or platforms.

The Free Primary Source Stack That Solo Practitioners Actually Use

Google Scholar provides federal and state court opinions with citation history, giving you a searchable case law database at no cost. Cornell's Legal Information Institute, integrated directly into OpenCase, delivers the full US Code and the Code of Federal Regulations, organized by topic and updated regularly, without a subscription.

The Federal Register integration in OpenCase extends that statutory and regulatory layer to daily rulemaking updates, so you are working from current regulatory text rather than a cached version of uncertain age. Official court websites and state legislature portals round out the stack for local rules, procedural requirements, and session laws that neither platform always captures. OpenCase's daily PACER integration adds federal docket access to that foundation, surfacing filings and orders that case-law databases alone will not show you.

Library, Online, or AI - A Prioritized Answer to the Question Every Beginner Asks

Begin with a secondary source to orient yourself on the legal standard and identify the key authorities. From there, move to Cornell LII for the governing statute or regulation, then to Google Scholar for the cases your secondary source flagged. Official court and agency websites fill the gaps.

OpenCase addresses both pressure points directly. Its Microsoft Word and Google Docs integrations let you run statute analysis, file analysis, and document drafting without leaving the document you are already in.

Across the market, attorneys weighing these subscription costs consistently flag workflow efficiency, not raw database coverage, as the deciding factor. An AI legal research platform that searches across 100+ legal databases, integrates Cornell LII, Federal Register, and PACER, and delivers source-backed answers inside the tools you already use makes the free-source sequence a starting orientation rather than the entire research burden.

The Step-by-Step Legal Research Process for Finding Primary Sources

Legal research fails most often not at the search step, but at the sequence step. Researchers who jump straight to keyword searches in case law databases routinely miss the controlling statute sitting one layer above every case they find. The funnel runs in one direction: wide orientation first, narrow binding authority last. Here is the exact sequence.

Step 1 - Mine Secondary Sources First to Map the Legal Landscape

Before touching a single statute or case, experienced legal researchers open with secondary sources, treatises, law review articles, ALR annotations, and legal encyclopedias like Am. Jur. or C.J.S. These tools identify the controlling legal framework, surface leading case citations, and point directly to relevant statute and regulation numbers. The tradeoff: secondary sources are never citable authority, so this step is orientation only, not the destination.

Step 2 - Locate the Governing Statute or Regulation Before Searching Case Law

Once the framework is clear, move immediately to binding primary text: the governing statute in the official or annotated code, or the controlling regulation in the CFR or its state equivalent. Reading the enacted text first anchors every subsequent case search in the actual words courts must interpret. The common mistake is jumping straight to case law and missing a statutory amendment that changes the entire analysis.

Step 3 - Expand Outward Through Citation Networks, Not Just Keyword Searches

Once one strong on-point case is in hand, use Westlaw's KeyCite or Lexis's Shepard's to follow the chain of authority in both directions, cases that cited your case and cases your case relied upon. This citation-network method surfaces authorities that keyword searches routinely miss because courts use different terminology. The limitation: citation tools only cover published decisions, so unpublished or slip opinions require a separate targeted search.

Step 4 - Filter Aggressively by Jurisdiction and Court Hierarchy

Binding authority from the governing court, the highest court in the controlling jurisdiction, always trumps persuasive authority from other circuits or states, regardless of how well-reasoned the outside opinion is. Every case pulled during research should be tagged in the research log as binding or persuasive, with the court and date noted. Researchers who skip this step risk building arguments on persuasive authority while missing contrary binding precedent that opposing counsel will certainly find.

Step 5 - Validate Currency by Shepardizing or KeyCiting Every Primary Authority

A case that has been overruled, distinguished, or superseded by statute is worse than no authority at all, it actively harms the brief or memo. Before any primary source is relied upon, run it through a citator to confirm it remains good law. This step is non-negotiable even for well-known landmark cases, since subsequent legislative action can effectively nullify judicial holdings without formally overruling them.

Step 6 - Maintain a Running Research Log That Tracks Queries, Sources, and Authority Classification

A structured research log, recording every search query run, database used, results reviewed, and each authority's binding or persuasive status, prevents duplicated effort, supports supervision and review, and creates an auditable record if research decisions are later questioned. For paralegals and associates, the log also demonstrates thoroughness. The tradeoff is time investment upfront, but it consistently saves more time than it costs on any matter lasting more than a single session.

Step 7 - Synthesize Findings into a Coherent Authority Hierarchy Before Drafting

After gathering and validating all primary authorities, the final research step is synthesis: organizing sources into a clear hierarchy, controlling statute, binding case law interpreting it, then persuasive authority, before a single word of the memo or brief is written. This prevents the common error of drafting around whichever case was found last rather than the most authoritative. The limitation is that synthesis requires legal judgment, not just retrieval skill, making it the step most dependent on attorney oversight.

How to Verify You Are Using Good Law Before You Cite Anything

Retrieval and validation are two separate acts, and the gap between them is where sanctions live. Citation validation is the professional-responsibility checkpoint that determines whether research is competent or negligent.

Solo lawyer's hand over tablet with red warning flag invalidating a legal citation

What "Good Law" Actually Means and Why That Status Expires

As iPleaders' analysis of AI-generated citation risk documents, a source that was valid when decided may have been overruled, amended, or superseded by subsequent authority; its validity changes over time and must be actively re-confirmed before use.

A case decided in 2018 that supported your argument may have been quietly reversed in 2022, and without active re-confirmation, that shift is invisible until it becomes a problem in court.

How to Read a Citator and Which Flags Should Stop You Cold

Shepard's (on Lexis) and KeyCite (on Westlaw) are the two primary citator tools for citation validation in United States legal research, and understanding what their signals mean is the foundation of any competent verification workflow.

A red flag on KeyCite or a red stop signal on Shepard's means the case has been overruled or reversed on at least one point of law, and any citation to that authority requires immediate reassessment before the work product leaves your hands.

OpenCase's search spans 100+ legal databases, so the underlying opinion pool feeding that check is comprehensive. You are not confirming good law against a partial corpus and mistaking coverage gaps for clean signals.

Statutes and Regulations Need Verification Too

For federal statutes, OpenCase integrates directly with Cornell LII, giving you a current, cross-referenced view of statutory text without a separate browser workflow. For federal regulations, OpenCase's Federal Register Integration surfaces live amendment history alongside the regulatory text, so the version you are reading reflects what is actually in force, not a snapshot from the last time you manually checked. OpenCase's Daily PACER Integration extends that discipline into docketed filings, so recent court orders that modify or interpret a statute appear in the same research environment rather than surfacing only after you have already drafted.

Why AI-Generated Citations Fail Exactly at This Step

This is the precise failure point, and it is the one professionals we work with find most disorienting, because AI-generated citations look authoritative even when they are not.

iPleaders' reporting documents how a plausible-looking AI-generated citation that does not exist, or that misquotes a real case, creates professional and disciplinary risk for the citing lawyer regardless of how confident the AI tool appeared when producing it. Courts have already acted on this: WZZM13's coverage of DOJ lawyer sanctions documents a judge ordering a lawyer to show cause after AI-generated citations in federal filings could not be verified, a consequence that attached to the attorney, not the tool that produced the error.

OpenCase grounds every cited authority against CourtListener and its broader corpus of real opinions, so the citations the platform surfaces link to documents that actually exist and can be opened, read, and confirmed, rather than to generated text that approximates what a real opinion might look like.

Verification is the gate that separates retrieved information from citable authority, and it is also the step that exposes the sharpest limitation of general-purpose AI tools in a legal workflow, a limitation that purpose-built legal research platforms are specifically designed to close.

How Legal AI Research Tools Built for Lawyers Change This Process

Purpose-built legal AI platforms exist precisely because that verification handoff exposes a structural problem general tools cannot self-correct. Research puts the scale of it plainly: 94% of lawyers are already using AI, yet 83% don't trust it. Attorneys are using tools they distrust because nothing in those tools makes it easy to measure how grounded any given output actually is. Purpose-built legal AI makes uncertainty visible and measurable, replacing a binary fear of hallucination with a calibrated confidence signal attorneys can act on or escalate.

"Legal AI research tools face significant legal risk when built on copyrighted legal databases (e.g. Thomson Reuters), creating uncertainty about the viability of AI-powered legal research platforms."

— what we hear from legal AI startup founders

Generic AI shows broken citation; legal AI platform displays verified confidence signal

Why General AI Tools Break Down During Citation Verification

General-purpose AI generates output that looks authoritative but is not grounded in verified sources, so you get a paragraph that reads like a brief excerpt, complete with a case name and reporter citation, and the case does not exist. Industry analysis describes the hallucination problem as "only getting worse": output that looks like research but collapses at the exact moment verification is required.

What 'Purpose-Built for Legal Research' Actually Means

Unlike general tools trained on the open web, a legal AI research platform is trained on authoritative legal datasets. OpenCase, for example, integrates with Cornell LII and authenticated government sources, so when the platform surfaces a statute or opinion, the citation traces back to a primary source you can click through and verify yourself.

That design choice transforms the verification step from a hunt into a confirmation.

The question shifts from "does this case exist?" to whether the case supports the argument at hand, which is the question a lawyer should be spending time on.

Where Legal AI Tools Help in Your Research Process (And Where They Don't)

Purpose-built legal AI accelerates retrieval and surfaces cited authority so practitioners arrive at the verification step with real citations already in hand. What it cannot do is apply judgment: deciding which authority controls, reading the holding against your specific facts, or evaluating whether an adverse case distinguishes cleanly. Those analytical tasks remain with the attorney. As multiple industry sources note, AI makes legal research "easier, cheaper, and far more reliable," but not complete without the attorney's judgment on which authority controls and how it applies to specific facts.

Next steps

If your research hours keep ending in uncertainty about whether what you found will actually hold up in court, the path forward starts with treating verification as a required second phase, not an optional final check. Retrieval surfaces candidates. Confirmation determines whether any of them are safe to cite.

A structured research plan is a malpractice-prevention tool, not just an efficiency aid, which means skipping deliberate issue-framing and jurisdiction-locking creates professional liability exposure before you ever open a database. Citation validation is the professional-responsibility checkpoint that separates competent research from negligent research, which means unvalidated authority submitted to a court creates sanctions risk regardless of how strong the initial retrieval was. Together, they point to a single practical next step: run your research inside a tool that grounds every citation in verifiable primary authority and keeps the verification step inside the same workflow, not bolted on afterward.

Start with legal AI built specifically for this process. OpenCase searches across 100+ legal databases, integrates Cornell LII and the Federal Register directly, and surfaces citations that trace back to documents you can open and confirm, so the question shifts from "does this case exist?" to whether it controls your argument.

Frequently Asked Questions

What is legal research, really, isn't it just finding the right case?

Legal research is chain-building, not source-collecting. The goal is a traceable line from your client's facts, through applicable statutes or regulations, to controlling case law in the right jurisdiction, ending at authority a court will accept. Finding a case is only the retrieval step; confirming it is still good law and has not been quietly overruled is a separate and equally critical verification step, and skipping it is where most research errors actually occur.

Why does legal research matter so much for malpractice risk?

According to the ABA Journal's coverage of legal malpractice trends, the leading drivers of professional liability are failure to know the law, failure to apply it correctly, and failure to investigate facts, not simply failure to locate a source. A structured research process that locks in jurisdiction, verifies authority, and logs sources is itself a malpractice-prevention tool, not just a time-management aid.

How should I keep track of what I've found during legal research?

Build a research log before your first query and update it as you go, a simple table with columns for source name, citation, jurisdiction, date accessed, and validation status works well. The log functions as both a duplication check and a diligence record, and it separates a defensible research file from a loose collection of links.

Can I do competent legal research without a Westlaw or LexisNexis subscription?

Yes, but sequencing matters. Start with secondary sources, treatises, practice guides, and legal encyclopedias, to get a doctrinal map and direct citations to controlling primary authority before you open any case-law database. Free primary-source resources exist, but their coverage of older state appellate decisions can be uneven, and their citation-history features are not substitutes for a formal citator, so you still need a structured verification step.

What are the risks of using a general AI chatbot for legal research?

General AI tools do not actually close the legal research access gap. LexisNexis's 2024 guidance on generative AI in legal research identifies fabricated or unverifiable citations as a risk category that is separate from, and in addition to, the access problem small-firm attorneys already face, meaning a solo attorney who asks a general chatbot for controlling authority may receive a citation that looks precise, reads persuasively, and does not exist.