There is a particular kind of panic that visits every first-year law student, every self-represented litigant, and every paralegal who has just been handed a stack of printouts and told to "see if this is still good." You have a case. It seems to say exactly what you need. The language is perfect, the facts are close, the judge sounds confident. And then a small, nagging voice asks: Is this case still the law? Did I read it correctly? Does this court even have to follow it? That voice is the beginning of competent legal research. This guide is about answering it.

Case law is law made by judges. When a court decides a dispute and explains its reasoning in a written opinion, that opinion does more than settle the quarrel between the two parties in front of the court. In the common-law tradition the United States inherited from England, a court's reasoning can bind future courts confronting similar questions. That is the engine of the entire system: a decision becomes a rule, the rule guides the next case, and the law grows case by case, like coral. To research case law well, you need to understand three things at once: where decisions come from (the court systems and the doctrine of precedent), what a decision actually says (the anatomy of an opinion, and the all-important difference between its holding and its asides), and how to find and verify it (reporters, citations, digests, databases, and citators).

This is the deeper, case-law-specific companion to our beginner's guide to legal research. If terms like "primary authority," "mandatory versus persuasive," or "the hierarchy of sources" are new to you, start there; this guide assumes you know that cases are primary authority (the law itself, as opposed to commentary about the law) and focuses on the craft of working with them. By the end you will be able to read a citation cold, tell a holding from a throwaway line, brief a case in a way that actually helps you, find the cases you need three different ways, and—most importantly—confirm that the case you are about to stake your argument on has not been quietly gutted by a later decision.

A quick orientation before we dive in. Throughout, we will use a handful of invented parties to keep examples concrete. Meet Acme Corp., a manufacturer that keeps getting sued; Beta LLC, a scrappy competitor; and Dana Plaintiff, an individual with a grievance. None of them are real, and any resemblance to your own litigation is the law working as intended.

What Case Law Is, and Where It Fits

Each branch of American government produces a different species of law. Legislatures (Congress, the state legislatures) enact statutes. Executive agencies, acting under authority delegated by statute, issue regulations and administrative decisions. And courts, in the course of deciding disputes, produce case law—also called common law, decisional law, or judge-made law. All three are primary authority. None of them exists in a vacuum: a great deal of modern case law consists of courts interpreting statutes and regulations, deciding what ambiguous words mean and how they apply to facts the legislature never anticipated.

That interpretive function is why case law matters even in heavily statutory fields. A statute might say an employer cannot discriminate "because of sex." It takes a court to decide whether that phrase covers discrimination against a transgender employee—and once the Supreme Court answered that question in Bostock v. Clayton County, 590 U.S. 644 (2020), the statute's words meant something they had not concretely meant before. The statute is the text; the case law is the text as construed by the courts that enforce it. You cannot understand one without the other. This is also why, when you research a statute, you are not finished until you have read the cases interpreting it—a point we return to when we discuss annotated codes.

The power of courts to say what the law is, and to bind even the other branches, traces back to Marbury v. Madison, 5 U.S. (1 Cranch) 137 (1803), in which Chief Justice John Marshall famously declared that "it is emphatically the province and duty of the judicial department to say what the law is." That sentence is the cornerstone of judicial review and, by extension, of why a single appellate opinion can ripple across thousands of later disputes. Case law is not advisory commentary. In its own jurisdiction, on its own issue, it is the law.

The American Court Systems

You cannot evaluate a case until you know which court decided it, because a court's place in the hierarchy determines how much weight its decision carries. The United States runs parallel court systems: one federal, fifty state systems, plus the District of Columbia and the territories. Both the federal and state systems are layered into trial courts and appellate courts, and the layering is the whole point.

Trial courts are where cases begin. A trial court (in the federal system, a United States District Court; in the states, variously called superior courts, circuit courts, courts of common pleas, or simply district courts) finds the facts—who did what, when, and with what intent—and applies the law to those facts. Trials produce witnesses, exhibits, verdicts, and judgments. They also produce a great deal of paper; if you want to understand the documents generated along the way, our overview of evaluating and assessing a civil case walks through how practitioners size up a dispute from the trial-court trenches.

Appellate courts do something different. They do not retry the case or hear new witnesses. They review the trial court's work to make sure the law was applied correctly, examining the cold record for legal error. Because appellate courts decide questions of law for whole categories of future cases—not just the dispute in front of them—their opinions are where most binding precedent is made.

The Federal Hierarchy and Precedent

The federal system has three tiers, and the precedential relationships among them are clean enough to memorize.

At the top sits the Supreme Court of the United States, a single court whose decisions bind every other court in the country on questions of federal law and the federal Constitution. When the Supreme Court speaks on the meaning of the Constitution or a federal statute, every federal and state court must follow. There is no appeal from the Supreme Court; the only thing that can change a constitutional ruling is the Court itself overruling its earlier decision, or a constitutional amendment.

In the middle are the United States Courts of Appeals, organized into thirteen circuits: twelve regional circuits (eleven numbered circuits plus the D.C. Circuit) and the Federal Circuit, which hears specialized appeals such as patent cases and claims against the United States. A circuit's published decisions are binding precedent on every district court within that circuit—but only within that circuit. The Fifth Circuit's ruling binds district courts in Texas, Louisiana, and Mississippi; it does not bind a district court in New York, where the Second Circuit governs. A decision from another circuit is persuasive but not mandatory: a court may follow it because the reasoning is sound, but it is not obligated to. This is why the same federal statute can mean one thing in California and a different thing in Georgia until the Supreme Court steps in to resolve the disagreement—a "circuit split" that is itself a common reason the Supreme Court agrees to hear a case.

At the base are the United States District Courts: ninety-four districts, each with a district court (and an associated bankruptcy court), supplemented by specialized tribunals such as the Court of International Trade and the Court of Federal Claims. District courts must obey the Supreme Court and the court of appeals for their own circuit. Critically, a district court's own decisions do not bind other district courts—or even other judges in the same district. A district-court opinion is persuasive authority at best. Two judges in the same courthouse can reach opposite conclusions on the same legal question, and neither is "wrong" in the precedential sense until a higher court resolves the issue.

The State Hierarchies

Each state runs its own version of this pyramid, and the structure is broadly similar even when the names differ. Most states have a trial level, an intermediate appellate court, and a court of last resort. The court of last resort is usually called the "Supreme Court," but watch for traps: in New York, the trial-level court of general jurisdiction is confusingly named the "Supreme Court," and the highest court is the Court of Appeals. In Maryland and a few others, naming conventions have shifted in recent years. The lesson is to identify a court by its function and position in the hierarchy, not by assuming its name tells you the whole story.

The precedential logic mirrors the federal system: a state's highest court binds all lower courts in that state on questions of that state's law, and the intermediate appellate courts bind the trial courts below them. A decision from another state is, like an out-of-circuit federal decision, merely persuasive.

Where the Two Systems Meet

State and federal courts are not hermetically sealed. Federal courts routinely decide questions of state law (for example, in "diversity" cases between citizens of different states), and when they do, they must apply state law as the state's own highest court would—a rule announced in Erie Railroad Co. v. Tompkins, 304 U.S. 64 (1938). Conversely, state courts regularly decide federal questions and are bound by the U.S. Supreme Court's interpretations of federal law. So when you find a federal district court opinion predicting how the Ohio Supreme Court would rule on an Ohio contract question, remember that the federal court's prediction is not binding on Ohio courts at all; the Ohio Supreme Court remains the final word on Ohio law. Knowing whose law a court is applying, and whether that court is the authoritative voice on that law, is half of evaluating a precedent.

Stare Decisis: The Doctrine of Precedent

Everything above rests on one Latin phrase. Stare decisis—literally "to stand by things decided"—is, in the words of Black's Law Dictionary, the "doctrine of precedent, under which a court must follow earlier decisions when the same points arise again in litigation." It is the discipline that makes case law law rather than a series of unrelated opinions. Without it, every dispute would be decided from scratch, the outcome would depend entirely on which judge you drew, and no one could plan their affairs around any rule. Stare decisis trades a little flexibility for a great deal of predictability, stability, and equal treatment of like cases.

It comes in two flavors, and the distinction is worth internalizing.

Vertical stare decisis runs down the hierarchy. The decisions of a higher court create mandatory (binding) precedent that lower courts in the same jurisdiction must follow, whether they like it or not. A district judge who thinks her circuit's rule is wrong must apply it anyway; her remedy is to write an opinion explaining her disagreement and hope the court of appeals or Supreme Court eventually agrees. Vertical stare decisis is the non-negotiable spine of the system.

Horizontal stare decisis runs sideways through time: a court generally follows its own prior decisions when similar facts return. The Supreme Court usually follows its own precedents; a state's highest court usually follows its own. But horizontal stare decisis is a policy of consistency rather than an absolute command, and a court can overrule itself. When it does, it is a significant legal event. The Supreme Court overruled the "separate but equal" doctrine of Plessy v. Ferguson, 163 U.S. 537 (1896), in Brown v. Board of Education, 347 U.S. 483 (1954); it overruled Planned Parenthood v. Casey, 505 U.S. 833 (1992), and Roe v. Wade, 410 U.S. 113 (1973), in Dobbs v. Jackson Women's Health Organization, 597 U.S. 215 (2022). Each of those reversals reset the law in its area overnight and instantly turned a mountain of older case law into history rather than authority. This is precisely why verifying that a case is still good law is not optional—a point we hammer home later.

Courts do not overturn precedent lightly. They weigh factors such as whether the old rule has proven unworkable, whether people have come to rely on it, whether the legal landscape around it has eroded, and whether it was poorly reasoned to begin with. As a researcher, you should treat a long-settled, often-cited rule as far more durable than a fresh, fractured, or widely criticized one—but you should never assume any precedent is immortal.

A final nuance: the binding force of a case attaches only to its holding, not to everything the court happened to say. That distinction is so important it gets its own section. But first, you have to be able to read the opinion the holding lives in.

The Anatomy of a Judicial Opinion

Open a published appellate decision and, once you learn the parts, it reads like a well-organized argument. The components recur across courts.

The caption names the parties (Dana Plaintiff v. Acme Corp.), the court, the docket number, and the date. On appeal, the party who lost below and brought the appeal is the appellant (or, for a discretionary review, the petitioner); the other side is the appellee (or respondent). Note that a case can be recaptioned on appeal, so the first-named party in the citation may be the original defendant. Do not assume the first name is the plaintiff.

A synopsis or summary, usually written by a commercial publisher rather than the court, gives a quick overview of the dispute and the outcome. It is convenient but not part of the official opinion—you may not cite it as authority, only the court's actual words.

Headnotes follow in many published versions. These are editorial summaries of each discrete point of law the case decides, drafted by the publisher (West, LexisNexis) and numbered. They are research gold for finding related cases (more on West's Key Number System below) but, like the synopsis, are editorial matter. Never quote a headnote as if it were the court's holding; always go to the corresponding passage in the opinion itself.

Then comes the opinion proper, typically structured as: a statement of the facts and the procedural history (what happened in the lower courts and how the case arrived here); the issue or issues presented; the court's analysis (the legal reasoning, weaving statutes, prior cases, and policy); and the disposition (affirmed, reversed, vacated, remanded, or some combination). "Affirmed" means the appellate court agreed with the result below; "reversed" means it overturned it; "vacated" wipes out the lower ruling; "remanded" sends the case back for further proceedings consistent with the appellate opinion.

Majority, Concurrence, and Dissent

On a multi-judge appellate court, not everyone always agrees, and the opinion may come in several voices.

The majority opinion is the controlling opinion—the one joined by more than half the judges who heard the case. Its reasoning and result are the precedent. When a majority cannot agree on the reasoning but agrees on the result, you get a plurality opinion: the largest bloc's reasoning, which carries less precedential force and can require careful analysis to extract a binding rule (the Supreme Court's guidance in Marks v. United States, 430 U.S. 188 (1977), tells lower courts to find the narrowest ground commanding a majority).

A concurrence is written by a judge who agrees with the majority's result but wants to add or qualify reasoning. Concurrences are not binding, but an influential concurrence can become the seed of future law—Justice Brandeis's and Justice Harlan's concurrences have shaped whole doctrines over time.

A dissent is written by a judge who disagrees with the result. A dissent has no precedential force—it is, by definition, the losing view. Yet dissents matter to researchers for two reasons. First, they often sharpen the majority's holding by attacking it, helping you see exactly what the majority did and did not decide. Second, dissents sometimes age into majority rules; Justice Harlan's lone dissent in Plessy ("Our Constitution is color-blind") was vindicated decades later in Brown. When you cite a dissent, label it as a dissent—citing it as if it were the holding is a credibility-destroying error.

Holding Versus Dicta: The Most Important Distinction in Case Law

Here is where competent researchers separate themselves from the rest. A judicial opinion is full of sentences. Only some of them are law.

The holding (closely related to the civil-law concept of ratio decidendi, "the reason for the decision") is the court's resolution of the legal issue actually presented, as applied to the facts before it. It is the part of the opinion that was necessary to decide the case. The holding is what binds future courts under stare decisis. Everything else is, in varying degrees, commentary.

Dictum (plural dicta, from obiter dictum, "a thing said in passing") is everything the court says that was not necessary to the result: hypotheticals, observations about how the law might apply in different circumstances, musings about policy, answers to questions the parties did not raise. Dicta can be thoughtful, even prophetic, and lawyers cite it all the time—but it is not binding. A later court is free to disregard dicta entirely. Chief Justice Marshall said as much in Cohens v. Virginia, 19 U.S. (6 Wheat.) 264, 399 (1821), warning that general expressions in an opinion are to be read in connection with the case actually decided, and that broad language "ought not to control the judgment in a subsequent suit when the very point is presented for decision."

Why does this matter so much in practice? Because a perfectly worded sentence that looks like it resolves your problem may be dictum that no court is required to follow. Consider a worked example.

Suppose Dana Plaintiff sues Acme Corp. for breach of a software license, and the only issue the court must decide is whether the contract's notice provision was satisfied. The court holds that Dana gave proper notice and rules for Dana. Along the way, the judge writes a paragraph musing that "had Acme included an arbitration clause, this dispute would surely have been sent to arbitration." That arbitration sentence is dicta—the case had no arbitration clause, the question was never presented, and the observation was not necessary to the result. The holding is narrow: on these facts, the notice provision was satisfied. If, a year later, Beta LLC tries to cite Dana v. Acme for the proposition that "courts will enforce arbitration clauses in software licenses," Beta is leaning on dictum, and opposing counsel will pounce. (If Beta actually needs authority on that point, it should look to genuine arbitration precedent and to resources like our guide on arbitration as alternative dispute resolution.)

Distinguishing holding from dicta is partly art. Courts and commentators disagree at the margins, and an opinion's "holding" can be stated narrowly or broadly. As a working method: ask what question the court had to answer to reach its result, identify the facts that were essential to that answer, and treat the rest as persuasive-at-best. When you read a string of cases, you are really hunting for each one's holding and then asking whether your facts fall inside or outside it. That hunt—"is my case more like the binding holding or distinguishable from it?"—is the daily work of legal argument.

Published Versus Unpublished Decisions

Not every decision a court issues carries the same weight, and the labels can mislead.

In the federal courts of appeals, decisions selected for publication appear in the Federal Reporter and are binding precedent on the district courts in that circuit. Decisions not selected for publication—"unpublished" opinions—are collected (since 2001) in the Federal Appendix and generally carry no binding precedential force, though Federal Rule of Appellate Procedure 32.1 now permits parties to cite unpublished federal decisions issued after January 1, 2007. "Unpublished" is something of a misnomer in the internet age: these opinions are freely available online; they simply lack the precedential status of published ones. Each circuit has its own local rule on how much weight, if any, its unpublished decisions deserve, so check the rules of the specific circuit.

At the district court level, the picture is different again. District court opinions—whether printed in the Federal Supplement or not—are never binding precedent, because a trial court does not bind other trial courts. Published district court opinions tend to be treated as more persuasive than unpublished ones, but both are persuasive only. A beautifully reasoned district court opinion can be ignored by the judge down the hall.

States vary widely. Some give unpublished state appellate opinions no precedential value and even restrict citing them; others permit citation; a handful treat them as binding. Before you rely on an unpublished state decision, find that state's citation rule. The safest habit: prefer published, higher-court authority whenever you can get it, and treat unpublished and trial-level decisions as supporting color rather than load-bearing structure.

Reporters: Where Opinions Live

A reporter is a serial publication—originally bound books, now mostly databases—that prints judicial opinions in the order they are decided. Reporters group cases by court level: Supreme Court decisions go in one set, federal appellate decisions in another, and so on. Each time enough new decisions accumulate, a new numbered volume is issued. Understanding the reporter landscape is what lets you decode a citation and find the case.

Federal Reporters

U.S. Supreme Court decisions appear in three parallel reporters:

  • United States Reports (U.S.) — the official reporter published by the government. This is the citation form you should use whenever a U.S. cite is available.
  • Supreme Court Reporter (S. Ct.) — West's unofficial reporter, useful because it publishes faster and carries West headnotes and Key Numbers.
  • United States Supreme Court Reports, Lawyers' Edition (L. Ed., L. Ed. 2d) — LexisNexis's unofficial reporter, with its own editorial enhancements.

Because the same Supreme Court opinion appears in all three, it has three parallel citations. The official U.S. cite is preferred; the others are alternatives. (When a very recent decision has not yet been paginated in the bound U.S. volume, you will see a placeholder like "600 U.S. ___" with the S. Ct. cite supplied alongside.)

Federal courts of appeals decisions selected for publication go in the Federal Reporter, now in its fourth series: cited F., F.2d, F.3d, and F.4th as the series have rolled over. Published circuit decisions bind the district courts in that circuit. Unpublished circuit decisions (2001 onward) appear in the Federal Appendix (F. App'x) and generally lack precedential force.

Federal district court decisions, when published, appear in the Federal Supplement (F. Supp., F. Supp. 2d, F. Supp. 3d). Specialized reporters cover niche areas—Federal Rules Decisions (F.R.D.) for procedural rulings, the Bankruptcy Reporter (B.R.), and so on. None of these district-level decisions is binding precedent.

State and Regional Reporters

State decisions are published in two kinds of reporters, and the same opinion often appears in both.

Some states publish an official state reporter (for example, Virginia Reports for the Virginia Supreme Court). Others have abandoned official reporters entirely and designate a commercial reporter as official, or rely on the regional reporters.

West's National Reporter System divides the country into seven regions and publishes a regional reporter for each, gathering appellate decisions from the states in that region:

  • Atlantic Reporter (A., A.2d, A.3d): Connecticut, Delaware, D.C., Maine, Maryland, New Hampshire, New Jersey, Pennsylvania, Rhode Island, Vermont.
  • North Eastern Reporter (N.E., N.E.2d, N.E.3d): Illinois, Indiana, Massachusetts, New York, Ohio.
  • North Western Reporter (N.W., N.W.2d, N.W.3d): Iowa, Michigan, Minnesota, Nebraska, North Dakota, South Dakota, Wisconsin.
  • Pacific Reporter (P., P.2d, P.3d): Alaska, Arizona, California, Colorado, Hawaii, Idaho, Kansas, Montana, Nevada, New Mexico, Oklahoma, Oregon, Utah, Washington, Wyoming.
  • South Eastern Reporter (S.E., S.E.2d): Georgia, North Carolina, South Carolina, Virginia, West Virginia.
  • South Western Reporter (S.W., S.W.2d, S.W.3d): Arkansas, Kentucky, Missouri, Tennessee, Texas.
  • Southern Reporter (So., So. 2d, So. 3d): Alabama, Florida, Louisiana, Mississippi.

California and New York generate so much litigation that West publishes single-state reporters for them as well (the California Reporter, Cal. Rptr., and New York Supplement, N.Y.S.). When a state decision appears in both an official reporter and a regional reporter, you may need to provide parallel citations, depending on the rules of the court you are filing in. The Bluebook's Table T1 tells you which reporter each jurisdiction prefers and whether parallel cites are required—a small detail that local clerks enforce strictly.

A Note on Neutral and Vendor-Neutral Citations

Because so much law now lives online rather than in books, many courts have adopted public-domain or "neutral" citations that do not depend on a commercial reporter's pagination—often something like 2025 WI 14 (a Wisconsin Supreme Court decision) with numbered paragraphs instead of page numbers. These let anyone pinpoint a passage without owning the West volume. When a jurisdiction uses neutral citations, the Bluebook and that jurisdiction's rules will tell you to include them, frequently alongside the traditional reporter cite.

How to Read a Citation

A case citation is a compact set of coordinates for finding an opinion. Once you learn the pattern, you can read any of them at a glance. The standard elements are:

  1. The case name — the lead parties, italicized or underlined (Plaintiff v. Defendant).
  2. The volume number of the reporter.
  3. The abbreviation of the reporter.
  4. The first page on which the opinion begins (often followed by a pincite—a specific page you are quoting, e.g., "1013, 1017").
  5. A parenthetical with the court abbreviation (omitted for the U.S. Supreme Court, since the U.S. reporter implies it) and the year.

Take a real example: Stearns v. Ticketmaster Corp., 655 F.3d 1013 (9th Cir. 2011). Decoded, this tells you the case is between an appellant named Stearns and an appellee, Ticketmaster Corporation; the opinion appears in volume 655 of the Federal Reporter, Third Series, starting at page 1013; and it was decided by the United States Court of Appeals for the Ninth Circuit in 2011. From the citation alone you already know the court (a federal court of appeals), the level (intermediate appellate), and therefore the precedential reach (binding on district courts within the Ninth Circuit, persuasive elsewhere). That is an enormous amount of information packed into one line, and reading it fluently is a core research skill.

A Supreme Court example—Bostock v. Clayton County, 590 U.S. 644 (2020)—needs no court abbreviation because the U.S. reporter can only mean the Supreme Court. A state example with a parallel cite might read People v. Acme Corp., 12 Cal. 5th 100, 480 P.3d 1 (2021), giving both the official California reporter and the regional Pacific Reporter.

The Bluebook and Its Alternatives

The dominant citation manual in the United States is The Bluebook: A Uniform System of Citation, compiled by the editors of the Columbia, Harvard, University of Pennsylvania, and Yale law reviews. It is exhaustive, occasionally maddening, and the de facto standard in most courts and law schools. Some jurisdictions instead follow the ALWD Guide to Legal Citation or a state-specific style manual; a number of state courts publish their own citation rules that override the Bluebook for filings in that state. Always check local rules before assuming Bluebook governs.

For everyday purposes, two free resources are invaluable and worth bookmarking: Cornell Law School's Introduction to Basic Legal Citation (the Peter W. Martin guide at law.cornell.edu/citation) and the Georgetown Law Library Bluebook Guide. Treat them as supplements to, not substitutes for, the Bluebook itself when precision counts. And remember the broader point from our beginner's legal research guide: a citation is not decoration. It exists so that a judge, an opponent, or a future researcher can find your source, check your work, and trust your argument.

How to Read and Brief a Case

Finding a case is only the beginning; you have to extract its meaning. The time-tested method is the case brief—not the brief you file with a court, but a short, structured summary you write for yourself to capture what a decision actually decided. A good brief forces you to separate holding from dicta, to articulate the rule precisely, and to see how the case fits your problem. Generations of lawyers have used a version of the IRAC framework (Issue, Rule, Application, Conclusion). A practical case brief contains:

Caption and citation. The case name and full citation, so you can find it again and cite it correctly. Note the court and year here—they determine the case's weight.

Facts. The legally significant facts—the ones the court relied on. Resist the urge to copy the whole narrative; capture what mattered to the outcome. Ask: which facts, if changed, would have changed the result?

Procedural history. How the case got here. Who sued whom, what the trial court did, what the intermediate court did, and what this court is reviewing. Procedural posture matters enormously: an appeal from a grant of summary judgment is reviewed differently than an appeal after a full trial, and the standard of review can decide the case.

Issue. The precise legal question the court had to answer, framed as a yes/no question tied to the facts ("Did Acme give legally sufficient notice under the contract when it emailed Dana rather than mailing a letter?").

Holding. The court's answer to that question—the rule of the case—stated as narrowly as the facts permit. This is the load-bearing element. Be ruthless about distinguishing the holding from surrounding dicta.

Reasoning. Why the court reached its holding: the statutes and precedents it relied on, the analogies it drew, the policy it invoked, and how it dealt with the losing arguments. This is the part that lets you predict how the rule extends to new facts.

Concurrences and dissents. A line or two on any separate opinions, noting points that might matter later or that reveal the contours of the majority's holding.

Disposition and significance. What the court ordered (affirmed, reversed, remanded) and why the case matters to your problem—how it helps or hurts your position.

Here is the discipline that makes briefing worth the effort: a brief is not a substitute for reading the case, but a record of your analysis of it. The act of compressing a forty-page opinion into a single page is where you discover what you actually understand and what you only think you understand. When you later assemble a string of cases into an argument, your briefs are the raw material—each one a holding you can deploy, distinguish, or stack with others. This is the same analytical muscle our guide on evaluating and assessing a civil case brings to bear when sizing up whether a claim is worth pursuing in the first place.

A word on reading efficiently: you will rarely read every word of every case. Skilled researchers triage. Read the synopsis and headnotes to decide whether a case is even relevant; if it is, find the passage addressing your issue (the headnotes point you there); read that passage and the surrounding analysis closely; and read the rest only if the case proves central. The goal is not to admire the prose but to extract the rule and decide whether it binds, persuades, or distinguishes away.

Finding Cases: Three Roads In

There is no single way to find the cases you need; there are several, and good researchers move fluidly among them. Think of three principal roads.

By Topic (Digests and Headnotes)

When you know the subject but not any specific case, you need a topical finding tool. The classic is the West Digest system. West (now Thomson Reuters) editors read published opinions, break them into discrete points of law, and write a headnote for each point. Every headnote is filed under West's Key Number System, a taxonomy that carves all of American law into roughly 400 topics, each subdivided into specific Key Numbers—on the order of 100,000 of them in total. The genius of the system is that the same Key Number is used across every jurisdiction's digest. Find the Key Number that captures your issue (using the digest's Descriptive-Word Index), and you can pull every case in the country classified under it—federal, regional, or state.

In practice today, you usually meet the Key Number System inside Westlaw rather than in bound digests: read one good case, find the headnote that states your point, click its Key Number, and you are looking at a curated list of every other case that decided the same point. LexisNexis offers an analogous system of headnotes and topic classifications (it does not publish a West-style digest, but its headnotes link to related authority). This "one good case to many" move is one of the fastest ways to build out the law on a narrow issue.

By Citation (You Already Have the Cite)

Sometimes you have a citation—from a brief, a treatise, a statute's annotations, or another opinion—and simply need to pull up the case. Drop the citation into Westlaw, Lexis, Bloomberg Law, Google Scholar, or CourtListener and read the opinion. Then exploit the case as a hub: the cases it cites (going backward in time) and the cases that later cite it (going forward) are a map of the doctrine. Following citations forward—"who has relied on this case, and did they treat it well?"—is also the gateway to verification, which we turn to next.

By Statute (Annotated Codes)

This road is underused by beginners and indispensable to professionals. When your question turns on a statute, do not read the bare statutory text and stop. Read an annotated code—the United States Code Annotated (U.S.C.A.) from West or United States Code Service (U.S.C.S.) from Lexis for federal statutes, and the equivalent annotated state codes. After each statutory section, the annotated code collects Notes of Decisions: short summaries of every case that has interpreted or applied that section, organized by sub-issue. This is the single most efficient way to find the cases construing a statute, because someone has already gathered them for you and sorted them by topic. Statutes and the cases interpreting them are two halves of one whole; the annotated code is where they meet. (For the bigger picture of how statutes, regulations, and cases interlock, see our beginner's guide to legal research.)

A fourth, increasingly common road deserves mention: secondary sources as finding tools. A good treatise, an American Law Reports (A.L.R.) annotation, a law-review article, or a Practical Law practice note will not only explain an area of law but will collect and cite the leading cases—doing the topical search for you and adding expert commentary on which cases matter and why. Starting with a secondary source is often the fastest way into an unfamiliar area; the cases it cites become your entry points, which you then verify and expand using citators and digests.

Where to Find Cases: Free and Paid Tools

You can do excellent case-law research without spending a dollar, and you can also pay handsomely for speed, depth, and reliability. Knowing what each tier offers lets you match the tool to the task.

Free Resources

CourtListener (courtlistener.com), run by the nonprofit Free Law Project, hosts millions of opinions, oral-argument audio, and—through the RECAP Archive—a vast collection of federal court filings pulled from PACER. It is full-text searchable, free, and surprisingly powerful; it even offers citation tools to see how cases connect. For many practitioners and most members of the public, it is the first stop.

Google Scholar has a "Case law" search mode that covers federal and state appellate opinions, with "How Cited" links that show later citing cases. It is fast, familiar, and free, though its citator function is far less rigorous than the paid services.

govinfo.gov, the U.S. Government Publishing Office's portal, provides authenticated federal court opinions from selected courts (roughly 2004 to the present) and the official text of statutes and regulations. The Supreme Court's own website (supremecourt.gov) posts every opinion the day it issues, plus bound-volume PDFs back to 1991, and the Library of Congress has digitized older United States Reports volumes. The Administrative Office of the U.S. Courts maintains a Federal Court Finder so you can locate and visit any individual federal court's website, many of which post recent decisions.

Cornell's Legal Information Institute (LII) at law.cornell.edu offers free statutes, regulations, and a curated collection of significant opinions, all with helpful explanatory material—an excellent on-ramp for non-lawyers.

State court websites increasingly post their own opinions for free, and many state law libraries provide public research terminals. The Law Library of Congress's Guide to Law Online is a well-maintained directory of these free state and federal repositories.

Paid Subscription Services

Westlaw, LexisNexis, and Bloomberg Law are the comprehensive commercial platforms. They offer near-total coverage of federal and state case law, sophisticated full-text and natural-language search, the editorial enhancements discussed above (West headnotes and Key Numbers; Lexis headnotes), extensive secondary sources, and—critically—the gold-standard citators (KeyCite and Shepard's) that we discuss next. Their depth and reliability are why most firms and courts rely on them despite the cost. Bloomberg Law is especially strong on dockets and court filings. Thomson Reuters Practical Law layers practice notes, standard documents, and checklists on top of Westlaw, giving you expert-written explanations that double as case-finding tools.

If you cannot afford a subscription, do not despair. Many public law libraries, county law libraries, and academic law libraries provide free on-site access to Westlaw, Lexis, or Bloomberg "patron access" terminals. The Law Library of Congress, for instance, offers patron access to all three in its reading room. For most questions, a disciplined researcher can get very far on free tools and reserve the paid platforms for the verification step, where their citators are genuinely unmatched.

Verifying Good Law: Citators, Shepardizing, and KeyCite

We now arrive at the step you skip at your peril. Finding a case that says what you want is not the same as finding a case you can safely rely on. A decision may have been reversed on appeal, overruled by a later decision of the same or a higher court, vacated, abrogated by statute, limited to its facts, or simply criticized into irrelevance by a chorus of later courts. Cite a dead case to a judge, and at best you lose credibility; at worst, depending on the jurisdiction and how egregious the error, you risk sanctions for misleading the court. This is non-negotiable: before you rely on any case, you must confirm it is still good law.

The tool for this job is a citator—a system that tracks the subsequent history of every case and tells you how later courts have treated it. The two dominant citators are Shepard's (on LexisNexis) and KeyCite (on Westlaw); the verb "to Shepardize" (from the original Frank Shepard's printed citation books, which lawyers used for over a century before the systems went electronic) has become a generic term for citation-checking on any platform.

A citator does two distinct things, and you need both.

First, it reports the case's direct history: what happened to this very case as it moved through the courts. Was the decision you are reading affirmed or reversed on appeal? Was it later vacated or modified? If the Ninth Circuit opinion you love was reversed by the Supreme Court, you need to know that now, not when opposing counsel points it out.

Second, it reports the case's citing references (indirect history): every later case, statute, and secondary source that has cited it, with treatment signals indicating how it was cited. The platforms use color-coded flags or symbols—a red flag/stop sign on Westlaw or Lexis typically warns that the case has been reversed, overruled, or has serious negative treatment on at least one point; a yellow flag/triangle signals caution, such as criticism or partial negative treatment; a green or blue signal indicates positive or neutral treatment. Do not stop at the flag. A red flag may relate to a point of law you do not care about, while the holding you need remains perfectly good; conversely, a case with no flag at all can still be the wrong authority for your facts. The flag is a prompt to read, not a verdict. Click through to the negative-treatment cases and read them to learn what was undermined and whether it touches your issue.

Here is the verification workflow in plain steps. Pull up your case in a citator. Check the direct history first—if it was reversed or vacated, you may be done (find a better case). If the direct history is clean, scan the negative citing references: has any later court overruled it, abrogated it, declined to follow it, or limited it? For each negative treatment, read enough to know whether it affects your point of law or some unrelated holding in the same opinion. Finally, glance at the positive and neutral citing references to confirm the rule is alive and to harvest additional supporting cases. Only after this do you cite with confidence.

A free-tools caveat worth stating plainly: CourtListener and Google Scholar will show you that later cases cite your case (Google's "How Cited," CourtListener's citation network), which is genuinely useful, but they do not editorially classify the treatment the way Shepard's and KeyCite do. If your matter is significant—a brief to a court, an opinion letter, advice a client will rely on—use a true citator on a paid platform, or get to a library that provides one. For casual research, the free citation networks plus your own careful reading of the most recent citing cases will catch most problems. But never represent to a court that a case is good law unless you have actually checked.

The stakes here are not academic. Recall Dobbs v. Jackson Women's Health Organization, 597 U.S. 215 (2022): the day it issued, every brief and treatise that relied on Roe and Casey for a constitutional right to abortion became wrong. A researcher who pulled an old, persuasive-looking opinion built on Casey the following week, without running it through a citator, would have been citing a foundation that no longer existed. Citators exist precisely to catch these earthquakes—and the dozens of smaller tremors that reshape the law every term.

Dockets and Court Filings: The Paper Behind the Opinions

Opinions are the visible tip of litigation; beneath them lies a vast record of filings. A docket is, in the Administrative Office's words, a "log containing the complete history of each case in the form of brief chronological entries summarizing the court proceedings." Every case has a unique docket number, and the docket lists the parties, the dates of every appearance, and a one-line summary of each document filed—complaints, answers, motions, briefs, orders, and the rest.

Dockets and the underlying filings are not case law—they have no precedential value—but they are invaluable context. Reading the briefs the parties filed can illuminate why a court decided as it did; tracking a pending case's docket tells you whether a precedent you are relying on is itself on appeal and at risk; and the filings in a related case may contain the very evidence or argument you need. They are also where you go to authenticate documents and to understand the procedural machinery of a dispute. (If your matter involves digital evidence—screenshots, web captures, and the like—our guide on authenticating website screenshots as evidence in federal court explains how courts decide whether to trust such materials in the first place.)

For federal dockets and filings, the official source is PACER (Public Access to Court Electronic Records), the judiciary's electronic-records service. PACER is comprehensive but requires registration and charges per-page fees (with a modest free-usage threshold). The nonprofit RECAP Archive at CourtListener mirrors a large and growing share of PACER documents for free; whenever someone with the RECAP browser extension downloads a federal filing from PACER, it is contributed to the public archive. RECAP does not have everything, but it has a great deal, and it costs nothing—check RECAP before paying PACER.

For state dockets, access varies enormously from state to state and even county to county; the National Center for State Courts maintains a directory of state and local courts that provide online access. Among paid platforms, Bloomberg Law offers the most comprehensive docket coverage, and Lexis's Nexis Uni carries a collection of federal briefs and motions. A standing caution: documents posted on third-party blogs and aggregators are convenient but not authoritative—always verify a critical filing against the official court record before relying on it.

Putting It All Together: A Worked Research Session

Let us run a realistic problem end to end. Dana Plaintiff, a freelance developer, believes Acme Corp. used her copyrighted code without permission. You need to know whether a particular legal theory will fly in the Ninth Circuit (where Dana would sue). Here is how the pieces we have discussed fit together.

Start with a secondary source to frame the question and find entry-point cases. A treatise on copyright or a Practical Law note on software protection—see our overview of the legal protection of software—will explain the doctrine and cite the leading decisions. You harvest two or three case names.

Pull those cases by citation on CourtListener or Westlaw and read them, briefing each one: caption, facts, procedural history, issue, holding, reasoning. As you read, you separate each case's holding (binding) from its dicta (persuasive at best), and you note the court and year to gauge weight—a published Ninth Circuit decision binds the district court Dana would file in; an out-of-circuit decision is merely persuasive.

Expand by topic using the headnotes. One of your cases has a headnote stating exactly your point; you click its West Key Number and pull every Ninth Circuit case classified under it, plus persuasive authority from elsewhere. You also check the annotated code—the Notes of Decisions following the relevant section of the Copyright Act, 17 U.S.C.—to catch any statute-interpreting cases the digest missed.

Verify good law. For each case you intend to rely on, you run it through a citator (KeyCite or Shepard's if you have access; CourtListener's and Google Scholar's citation networks if you do not). You check direct history (was it reversed or vacated?) and negative treatment (has any later case overruled, abrogated, or limited it on your point?). You discard a promising older case because a recent Ninth Circuit decision limited it to its facts—exactly the kind of trap a citator exists to catch.

Assemble the argument. You now have a short stack of verified, on-point, binding (or strongly persuasive) authorities, each briefed, each holding pinned down, each confirmed as good law. You arrange them by weight (binding before persuasive, recent before stale, holdings before dicta), and you can articulate for each one how your facts fall inside its holding or outside an adverse one. That stack—and your honest assessment of the cases that cut against Dana—is the foundation of a sound legal position. Pairing it with a clear-eyed evaluation of the civil case as a whole tells Dana whether the fight is worth having.

Notice that no single tool did the whole job. Secondary sources framed the question and found seeds; citation lookup and briefing extracted the rules; digests and annotated codes filled out the field; and the citator guarded the gate. That choreography—frame, find, read, expand, verify—is the heart of case-law research, whether you are a Supreme Court advocate or a curious citizen.

Key Takeaways

Case law is judge-made law, and it is primary authority—the law itself, not commentary about it. To work with it confidently, hold these points in mind:

  • A court's place in the hierarchy determines its weight. Higher courts bind lower courts in the same jurisdiction (vertical stare decisis); out-of-jurisdiction and trial-level decisions are persuasive at best. Identify the court before you trust the rule.
  • Only the holding binds. The holding is the court's resolution of the issue it had to decide; everything else is dicta—often useful, never controlling. Learning to tell them apart is the central skill.
  • Read citations fluently. Volume, reporter, page, court, and year are all packed into a single line, and decoding it tells you the case's level and reach instantly.
  • There are many roads to the right cases: by topic (digests, headnotes, the Key Number System), by citation, by statute (annotated codes), and through secondary sources that collect cases for you.
  • Free tools go a long way. CourtListener, Google Scholar, govinfo, court websites, and Cornell's LII can handle most research; paid platforms (Westlaw, Lexis, Bloomberg Law, Practical Law) add depth, speed, and gold-standard citators.
  • Always verify good law. Use a citator (Shepard's, KeyCite, or at minimum a free citation network plus careful reading) to confirm your case has not been reversed, overruled, or undermined on the point you need. The flag is a prompt to read, not a verdict.

Master these, and the panic that opened this guide gives way to a calm, repeatable method. You will find the cases, read them for what they actually hold, and stand behind them because you checked.

Frequently Asked Questions

What is the difference between a holding and dicta, and why does it matter so much? The holding is the court's answer to the legal question it was required to decide, applied to the facts before it—it is the part of the opinion that binds future courts under stare decisis. Dicta are statements the court made that were not necessary to the result: hypotheticals, observations, and asides. A later court must follow a holding (if the deciding court is higher in the same jurisdiction) but is free to ignore dicta. It matters because a sentence that perfectly states your point may be unenforceable dictum, and an opponent who spots that will use it to dismantle your argument.

Are unpublished opinions worthless? No, but be careful. "Unpublished" no longer means "unavailable"—these opinions are online and citable in the federal courts of appeals (for decisions issued after January 1, 2007, under Federal Rule of Appellate Procedure 32.1). They simply lack binding precedential force and are treated as persuasive at most. State rules vary widely: some forbid citing unpublished decisions, others allow it. Always check the rules of the specific jurisdiction, and prefer published, higher-court authority when you can get it.

Do I have to use Westlaw or LexisNexis, or can I research case law for free? You can do excellent research for free. CourtListener, Google Scholar's case-law mode, govinfo, individual court websites, and Cornell's Legal Information Institute cover most needs. The one area where paid platforms remain genuinely superior is verification: KeyCite (Westlaw) and Shepard's (Lexis) classify how every later case treated your case, which the free citation networks do for free but with less editorial precision. For anything you will file or rely on professionally, run a true citator—and remember that many public and academic law libraries offer free on-site access to the paid platforms.

What does it mean to "Shepardize" a case? "Shepardizing" means checking a case in a citator to confirm it is still good law and to see how later authorities have treated it. The term comes from Frank Shepard's printed citation books, used for over a century before the function went electronic. Today it is a generic verb covering both Shepard's (Lexis) and KeyCite (Westlaw). The process has two parts: checking the case's direct history (was it affirmed, reversed, or vacated on appeal?) and its citing references (has any later case overruled, abrogated, or limited it, and on what point?). Skipping this step is one of the most dangerous mistakes a researcher can make.

How do I find cases that interpret a particular statute? Use an annotated code—the U.S.C.A. or U.S.C.S. for federal statutes, or the annotated version of the relevant state code. After each statutory section, an annotated code collects "Notes of Decisions": summaries of every case that has construed that section, sorted by sub-issue. This is far more efficient than reading the bare statutory text and guessing, because the cases have already been gathered and organized for you. Combine it with a secondary source and a digest search for thorough coverage.

What is the difference between binding and persuasive authority? Binding (mandatory) authority is precedent a court must follow: a decision from a higher court in the same jurisdiction on the same issue. Persuasive authority is precedent a court may follow if it finds the reasoning convincing but is not required to: decisions from other jurisdictions, lower or coordinate courts, dicta, dissents, and secondary sources. The same decision can be binding in one court and merely persuasive in another—a Ninth Circuit ruling binds district courts in the Ninth Circuit but only persuades a court in the Second Circuit. Our beginner's guide to legal research develops this distinction in detail.

Can I rely on the headnotes or the synopsis instead of reading the opinion? Use them to find and triage cases, never as a substitute for the court's actual words. Headnotes and synopses are written by commercial publishers (West, LexisNexis), not by the court, and they occasionally simplify or even misstate a point. They are excellent road maps—a headnote points you to the exact passage you care about, and its Key Number leads you to related cases—but the only thing you may quote or cite as the holding is the opinion itself.

Is a non-lawyer allowed to do this kind of research? Absolutely. Court opinions, statutes, and most research tools are public, and many of the best (CourtListener, Google Scholar, govinfo, Cornell LII) are free and designed to be usable by anyone. A self-represented litigant or a curious citizen can find and read the law. The cautions are the same ones professionals face: confirm a case is still good law before relying on it, distinguish holdings from dicta, and recognize that finding the law is not the same as knowing how a court will apply it to your particular facts. When the stakes are real, use this research to have a better-informed conversation with a qualified attorney—not to replace one.

Related Articles

This article provides general information for educational purposes only and is not legal advice. Reading and interpreting case law is fact- and jurisdiction-specific; for guidance on a particular matter, consult a qualified attorney licensed in the relevant jurisdiction.