The pace of litigation is dizzying. The path of every single lawsuit is filled with multiple inflection points, moments where attorneys have to make decisions about how the future is likely to unfold. In the past, attorneys navigated these twists and turns by relying on intuition, experience and anecdotal evidence. But things are starting to change. Judicial analytics remains one of the last frontiers of Big Data, a field poised to fundamentally transform the way attorneys practice the law by quantifying the unquantifiable to unimaginable ends.
Starting From Scratch
AI-powered judicial analytics emerged from the frustrations of day-to-day life as an attorney. Every attorney knows that the details of a past case can provide invaluable insights into how they should position similar cases in the future. The problem, however, was that these insights were impossible to access, especially for attorneys at the state trial court level. There was no effective way to perform practical legal research on state trial court records. The data was dispersed across thousands of separate courthouses throughout the country, with each county in each state authoring its own protocols for collecting, cataloging, and publishing court documents.
The result? State trial court was too scattered, too clunky and too inconsistent to prove useful.
Composed of dockets, petitions and rulings, state trial court data was designed for human (rather than mechanical) consumption. To be integrated into a series of interrelated data sets, this information needed to be cleaned, wrangled into standardized formats that a computer could read. This has been a difficult and expensive process. Consider, for a moment, names. The names of judges, attorneys and law firms are subject to change, a fact that makes it hard — if not impossible — to track legal entities across any given data set. Fortunately, much of this tedious work has already been performed by major legal technology platforms. Trellis Research, for example, has already mined trial court records across multiple states, taking all the content that goes into civil litigation (petitions, motions, rulings) and restructuring it so that it can be made searchable by a Google-like algorithm.
Current Capabilities of Judicial Analytics
It all starts with a judge dashboard. This is where attorneys can study an assigned judge’s caseload, with charts indicating the number of cases they have active and the average duration and type of practice area of those cases. These figures, culled from thousands of data points from court dockets, let attorneys get a sense of how fast cases move through a particular judge’s court. Knowing where a judge sits in relation to their peer averages, as well as the number of cases the judge has on deck, can help litigators anticipate the likely pace of their case.
Motions. Judicial analytics can also provide information about grant rates for different types of motions. The AI that guides these analytics platforms use a judge’s past rulings to model how the judge is likely to respond to motions filed by plaintiffs and defendants in the future. We can quickly see, for example, that the Hon. Kevin C. Brazile of the Los Angeles County Superior Court favors plaintiffs in labor and employment disputes and defendants in property disputes, granting 43% of demurrers for the former and 85% for the latter. This is crucial information to have when formulating settlement strategies. As one attorney from Locke Lord explained, “it [litigation analytics] also helps me predict, based on prior rulings, how the judge may rule on the motion. The information gives me a sense of who has leverage during settlement negotiations.”
Case outcomes. Judicial analytics can also offer information about case outcomes broken down by practice area. This is particularly helpful at the start of litigation when an attorney is tasked with deciding which court will be best for their client’s specific claim; an attorney can quickly get a sense of how local judges have handled similar cases in the past. Returning to Judge Brazile, we learn he has dismissed 233 (out of 290) of the wrongful termination cases assigned to him. He has, in contrast, presided over far fewer unlawful detainers, dismissing 36 (out of 53). This data can be used to assess a judge’s experience with different issues. If the judge presides over a specific type of issue often, an attorney can spend less time discussing the particulars of the law and move straight toward the facts of the case.
Milestones. And lastly, attorneys can perform milestone analyses. It takes only a second to learn that, on average, Judge Brazile cycles through property cases 43 days faster than his peers on the bench. This information is critical for the logistical side of litigation, helping litigators provide accurate timelines and budgets for their clients.
With judicial analytics, an attorney can begin to form a deep understanding of their judge’s tendencies. This deep understanding goes beyond surface-level anecdotes. “Judges are given broad discretion,” explains Donald Ricketts, a former civil litigation attorney in Southern California. “[And] if a particular ruling is discretionary, it is unlikely that an appellate court will reverse it.”
The Shift From Stories to Forecasts
“Before, when a motion was pending for a while, you were in the dark,” says Rusty Perdew, a litigation partner at Locke Lord. “You didn’t have insight into whether your situation was unusual or not. And you started guessing as to what that meant.” By synthesizing millions of unique data points, judicial analytics has uncovered the patterns nestled inside state trial court records, allowing attorneys to see things that nobody else can see. While this has proved useful for many litigators across the country, the technology of judicial analytics remains better at describing reality than it is at predicting it.
As we enter the coming year, we’ll see judicial analytic platforms begin to shift from description to prediction. “We are still largely describing, and counting, and presenting information in new and different ways,” says Jeff Arvidson, Director of Product Development at Thomas Reuters. “We can say, ‘Here are the trends over the last three years, five years, 10 years. These are the trends looking back.’ And either we leave you hanging right there, or eventually, we start to draw a dotted line forward to show you where that trend line goes.”
We’ll also see judicial analytics begin to explore more sophisticated statistical and theoretical techniques. More and more analytic platforms will incorporate advanced regression models to estimate the impacts of different variables on case outcomes. And we’ll see creative attempts to deploy computerized sentiment analysis of judicial rulings, a project that will produce more nuanced insights into individual judges’ ideologies and attitudes.
While We Wait: What’s Next for Judicial Analytics
Judicial analytics is the science of drawing insights from large quantities of data. This is a new science, one that requires its practitioners to pay careful attention to the results they generate. It’s important to remember that every insight derived from judicial analytics needs to be interpreted before it can become meaningful. And so, when presenting figures to clients, attorneys should enter those conversations by creating shared understandings of key statistical concepts. They might, for example, remind clients of the extent to which correlation does not necessarily indicate causation. That any apparent correlations may be due to a third, unidentifiable variable, or an underlying bias within the data set.
As more attorneys begin to incorporate judicial analytics into their practice, it might be tempting to take its insights for granted, relying on charts and graphs without considering the wider context behind the scene. It’s easy to forget just how misleading data viewed in isolation can be. As such, attorneys should not be afraid to probe beneath the surface, asking legal technology platforms about the underlying data behind a particular statistic. After all, it’s only after clients and attorneys have fully grasped the limitations of judicial analytics that they can really explore its possibilities.