Legal Citator: Online Citator and Case Citator Tools Compared

The citators do not agree with each other

This is the finding that should change how you use these tools, and almost nobody outside law librarianship talks about it. Paul Hellyer, in "Evaluating Shepard's, KeyCite, and BCite for Case Validation Accuracy," 110 Law Library Journal 449 (2018), reviewed 357 citing relationships that at least one citator had labeled negative. He found that Shepard's and KeyCite each missed or mislabeled about one third of the negative citing relationships in that sample, and that BCite missed or mislabeled more than two thirds. He also found examples of the most serious kind of citator error in all three products.

Sit with that for a moment. These are the editorial citators, the ones with attorney editors and decades of institutional process behind them, and on the specific job of catching negative treatment they were each wrong a meaningful share of the time, in different places. A clean flag is evidence that no editor recorded a problem. It is not proof that no problem exists.

The practical consequence is not that citators are useless. It is that a citator result is a starting point rather than a verdict, and that the confidence people place in a green symbol is higher than the underlying accuracy justifies. If a case is load bearing in your brief, read the negative and cautionary citing opinions yourself rather than trusting the label on them. Our guide to <a href="/blog/westlaw-keycite-flags-vs-shepards">KeyCite flags and Shepard's signals</a> sets out what each specific symbol actually claims.

Is there a free legal citator?

There is free citation data, and there is no free editorial citator. The distinction matters. Google Scholar has a "How cited" tab that shows you decisions citing the case, and CourtListener from the nonprofit Free Law Project publishes Authorities and Cited By lists across more than nine million decisions from over two thousand courts. Both will tell you that later courts cited your case. Neither will tell you that any of them overruled it.

That gap is the entire product. Assembling the citing list is a data problem, and it has largely been solved in the open. Deciding what each citing opinion did to the original is an editorial and analytical problem, and that is what the paid citators sell. Anyone telling you they have a free replacement for Shepard's is describing the first half of the job.

For a solo or small firm the honest calculus is usually this: use the free sources to find and read the cases, and pay for something that classifies treatment, whether that is a full Westlaw or Lexis plan, Authority Check through your state bar, or a tool like this one. What you should not do is treat a citing list with no treatment analysis as if it were a good law check. We go into the coverage tradeoffs on our <a href="/alternatives/courtlistener">CourtListener comparison</a>.

Editorial, algorithmic, or AI: which citator model should you buy?

Editorial citators are the conservative choice and remain the standard a court will expect you to have used. You are paying for attorney editors reading citing opinions, plus statute coverage, plus the fact that KeyCite and Shepard's are what opposing counsel is using. Their weakness is price and the Hellyer accuracy finding, which applies to them as much as to anyone.

Algorithmic citators like Authority Check are the value option, and for many solos they are already paid for through bar dues without the member realizing it. They are honest about being pattern based, they cover cases rather than statutes, and they are a reasonable safety net rather than a full replacement for editorial analysis.

AI citators are the newest category and the one with the widest spread between marketing and evidence. Treat published accuracy numbers as vendor claims until an independent study replicates them, and prefer tools that show you the treating opinions rather than only a verdict, because a signal you cannot audit is a signal you cannot defend. That is the design decision behind our own <a href="/features/citation-check">citation check</a>: every treatment label links to the opinions that produced it, so the work of checking stays possible instead of being hidden behind a symbol.

Where the word Shepardize comes from, and why it is a trademark

In 1873 Frank Shepard was a salesman for a Chicago legal publisher, and he noticed how lawyers were coping with a body of law that was changing fast after the Civil War: they kept handwritten notes in the margins of their reports about which cases had been overruled or affirmed. Shepard started printing those relationships instead, as gummed labels listing the later cases that had cited a given decision, which a lawyer could paste straight into the margin of a reporter volume.

The refinement that turned a list into a citator came next. Shepard's began adding one letter codes showing why a case had been cited, marking whether the citing court had overruled, criticized, modified, or applied the earlier decision. Every colored flag on every research platform today descends from those letters. By the early twentieth century the Frank Shepard Company was binding the citations into the maroon volumes with gold spines that generations of lawyers learned on.

The ownership trail explains the trademark. Shepard's passed to McGraw Hill in 1966, then to Times Mirror and Reed Elsevier in 1996, and LexisNexis took full ownership in 1998. "Shepard's" and "Shepardize" are registered trademarks of RELX Inc., the parent of LexisNexis. So when a tool other than Lexis offers to Shepardize your case, read that as a generic description of citation checking rather than as access to Shepard's. The distinction is not pedantry, because the editorial work is the thing you are actually paying for, and no two vendors have done the same editorial work.