Data-Driven Case Valuation

Parties to lawsuits and their
lawyers inevitably find themselves pondering one overriding question: What is
our case really worth?

The answer drives every major
decision in a case. Whether to negotiate. How hard to negotiate. Whether to
accept an offer or go to trial. This article discusses a data-based way to help
make those decisions by combining a key research insight on jury decision
making and input from mock jurors.

Challenges to Case Valuation

Seasoned trial lawyers think they
know something about case valuation. And why wouldn’t they? They know their
facts—the good ones and the bad ones. They know the demographic characteristics
of the jury pool, the history of jury awards in the jurisdiction, and the
awards produced by juries in similar cases. This knowledge and experience
certainly provides some insight into what a case is worth.

But it’s difficult to accurately
value a case using just rules of thumb or mental models filtered through the
lens of experience. Real numbers certainly come into play (e.g., actual medical
bills), but the list of intangibles is often long. Pain and suffering, lost
income, future medical costs, decreased enjoyment of relationships, expected
life span—among others.

Jurors also differ in their
backgrounds, life experiences, attitudes, and values. Diversity among jurors
leads to variability in how they value cases. Reasonable minds can and will differ.

And sometimes jurors consider
extralegal factors they shouldn’t. Research shows they often consider things
like whether the two parties are insured, how much of any award will go to the
attorneys, what will remain after taxes, etc.

How is a litigator supposed to
account for all these things in valuing a case?

A Data-Based Method for Valuing a Case

It all starts with one very
important and reliable finding: The social process of “converting” a set of
juror award preferences into a final jury award is conservative in that groups
usually choose awards near the middle of their members’ preferences.  In fact, research shows the best predictor of
a jury’s award is the median award preferred by its members. Studies have found
the median member award (prior to deliberation) to be highly correlated with
the jury’s actual award—and more so than the mean member award.  So the median rule provides a good empirical
basis for predicting a jury’s award.

And, once we have data from mock
jurors, we can use the median rule to forecast the jury award in any case.

We do this by collecting data from
a large pool of mock jurors and randomly choosing a mock “jury” of six people
from the pool (or whatever jury size is used in the trial jurisdiction). We
then estimate the award that “jury” of six people would have arrived at if they
had deliberated by calculating the median award preferred by its members. This
award is noted, the mock jurors are returned to the pool, and the sampling
process is repeated many times.

This process creates a
distribution of potential jury awards based on the preferred awards of our mock
jurors. So, if we randomly draw out 1,000 different juries, we will have an
award distribution consisting of 1,000 estimated jury awards. But what use is
this?

Our potential award distribution
provides two very important pieces of information: (1) the most likely jury
award, and (2) the most likely range of potential awards. This award
distribution will be bell-shaped, with most awards in the middle but a few that
are extremely high or low. Most importantly, the mean of our potential award
distribution is statistically the most likely award if we picked one jury at
random out of the pool of mock jurors. And the closer a potential award is to
the mean of our award distribution, the more likely we would be to get that
award from a randomly selected jury.

And we can go even further. We can
use probability theory to construct a confidence interval (often 95%) around
the most likely award to get a sense of the range of jury awards we might see.
So, for example, we might learn that if we were to pull out one random jury,
its most likely award would be $1.7 million. And if we pulled out 100 juries at
random, 95 of their awards would be expected to fall between $1.3 million and
$2.1 million (i.e., a 95% confidence interval).

But why not just estimate awards
using the preferred awards of the individual mock jurors? The short answer is
that their distribution is very different from the mock juries’ award
distribution. The individuals’ distribution is much more skewed and much more
spread out. Predicting jury awards from individual data basically results in a
lot more prediction error.

Technology now makes it easy to
collect data from a large and diverse group of mock jurors from the trial
jurisdiction via the internet. We can present a concise summary of key facts
and arguments, even show pictures or video clips, then ask for judgments
related to liability and damages. And it can all be done for a fraction of a
case’s value in a fairly short period of time.

Of course, the output from mock
jurors is only as good as the inputs. The information given to the mock jurors
needs to be carefully assembled, vetted, and presented. This is where seasoned
trial counsel plays an important role. These folks must advise the research
team on the information likely to come into evidence, the manner in which it
will be presented at trial, and the arguments both sides will make about it.

Valuing a case is difficult. But
case-specific data can make it less so. Armed with a forecast of potential jury
awards and their likelihood, parties and their counsel can make better
decisions about settlement and trial strategy by better answering the
underlying question: What’s this case
really worth?