When I tell lawyers that I’m teaching the LegalRnD version of “Quantitative Analysis for Lawyers” next semester at MSU Law, I usually get one of three reactions:
- Wow, what a great class! I wish I was still in law school!
- Wow, what a great class! But I’m not good at math.
- Why teach lawyers quantitative analysis?
With all the talk about big data, forensic evidence in the courtroom, artificial intelligence, code, and robot lawyers, the value of quantitative training is becoming obvious. Many lawyers see opportunities to apply quantitative thinking in practice, especially at the intersection of law and technology. At the same time, data and artificial intelligence are transforming legal-service delivery. The challenge of exercising basic math skills in an introductory quantitative analysis class is nothing compared to the rewards from learning quantitative thinking.
But there remain far too many lawyers and law students–especially law students–who do not see the connection between quantitative thinking and the law. Why should law students take “Quantitative Analysis for Lawyers”? The better question is, “How can law students afford not to learn quantitative thinking?”
Lawyers must be knowledgeable consumers of quantitative information.
Statistics and scientific thinking play a tremendous role in many legal fields. Some have argued that it plays such a large role in certain fields–such as employment discrimination, products liability, torts, and evidence–that statistics should be mandatory in law school. (Yair Listokin, Why Statistics Should be Mandatory for Law Students, May 22, 2006.)
Lawyers deal with quantitative information in numerous settings:
- Working with (or against) experts on damages calculations, product-failure investigations, establishing causation for toxic torts, forensic evidence in criminal cases, and employment discrimination modeling.
- Statistical sampling for health care fraud and other claims.
- Algorithms for sentencing.
- Dealing with electronically stored information, technology assisted review, and other eDiscovery issues.
- Conducting data-driven transaction and case assessments.
- Making decisions and advocating with statistical arguments.
This is but a small set of examples. Federal Court of Appeals Judge Richard Posner has said that lawyers’ discomfort with math, science, and technology is “increasingly concerning, because of the extraordinary rate of scientific and other technological advances that figure increasingly in litigation.” That is, there is a growing need for quantitative analysis by lawyers.
Individual rights are at risk when lawyers fail to engage in rigorous quantitative thinking. For example, Federal Court of Appeals Judge Alex Kozinski recently said in an opinion piece that many have been convicted based on flawed forensic-evidence techniques. If more lawyers were trained in quantitative analysis, overeager expert witnesses and the “voodoo science” Judge Kozinski condemns would be exposed by defense attorneys and principled prosecutors.
Big data, algorithms, and artificial intelligence are everywhere, raising important legal, business, and ethical questions.
Big data, algorithms, machine learning, code and the like regularly appear in the headlines. The information revolution has created countless substantive legal issues at the intersection of law and technology. Can lawyers properly address these issues without some understanding of the underlying technologies?
For example, predictive algorithms are used to make decisions about creditworthiness, employment, housing, insurance, policing, and sentencing, just to name a few. Studies have uncovered discrimination by algorithms, challenging popular misconceptions that data-driven algorithms are bias free.
Any lawyer can vaguely condemn discrimination in algorithms. But how much value can lawyers add if they don’t understand the basics of algorithms, machine learning, and computer code? Lawyers with an understanding of the underlying technology will be more effective.
Over time, clients will expect more from their lawyers, if they do not already. An excellent example played out at the ABA Business Law Section annual meeting this last September. At a panel on “FinTech: Machine-Learning Algorithms and Compliance Challenges,” I counted about 100 lawyers in the room learning about algorithms, linear regression, logistic regression, overfitting, and machine learning. Why did lawyers attend? They need this knowledge to enable them to counsel their clients to utilize technology properly and assess related risks. They need to be able to speak the language to be effective.
Prepare for the technology-infused delivery of legal services in the 21st Century.
Data, algorithms, machine learning, expert systems, and artificial intelligence will play an increasingly important role in legal-service delivery. This begins with basics, such as establishing metrics to quantify value and assess the quality of legal services. As legal processes are standardized and best practices emerge, more work previously done by lawyers will be automated.
At the same time, hype about artificial intelligence and robots practicing law has gotten out of hand. In our LegalRnD version of “Quantitative Analysis for Lawyers,” students will learn about these technologies through hands-on problem sets and projects. Why is learning these technologies important? Because it helps lawyers identify what work is likely to be automated and which tasks lawyers should perform to create value for clients.
Likewise, a growing body of data is available for risk assessments and legal predictions. Clients also want to mine their own data to predict and detect misconduct, fraud, and harassment and prevent it from escalating. Lawyers with quantitative skills can help their clients employ these technologies proactively, to prevent problems and promote business interests. (For example, General Motors recently posted the position, “Data Scientist – Legal.” The list of responsibilities highlights some of the ways in which data and analytics could be used.)
Learning some code and statistical tools, such as R, Python, and Tableau, can help law students build a career.
To paraphrase Seth Godin, if you can identify an entity that won’t be able to live without you, then you can find a job. It is becoming extremely difficult for lawyers to differentiate themselves based on traditional substantive legal knowledge, much less for law students to do so. But there are many opportunities for law students to differentiate themselves in emerging substantive areas, such as at the intersection of law and technology. Some examples include predictive algorithms, blockchain, information privacy and security law, 3D printing, drones, and autonomous vehicles.
The ability to read code, use statistical tools, and understand how artificial intelligence works are examples of additional ways that lawyers can add value. Whether lawyers should learn to code has been a topic of debate. Often missing is discussion about the depth of coding knowledge.
Lawyers should not strive to be developers–that would take a tremendous investment of time. On the other hand, many lawyers seem to believe that writing code is nothing short of magic. It should not, and does not, need to be that way. I’ve spoken to many lawyers who have invested the time to learn a basic language like Python through one of many excellent free sites online. After five to ten focused hours, you can credibly say that you can write some code–it is no longer a mystery.
If we view code as a language, we see the value of learning about it. If lawyers cannot speak a language, if they do not know any vocabulary, they cannot communicate with clients and experts. For example, lawyers who know nothing about finance struggle to communicate with business clients and accountants. Similarly, lawyers who know nothing about code will struggle to communicate with clients, data scientists, and developers.
An example of this arose at the ABA Business Law Section panel that I referenced above, “FinTech: Machine-Learning Algorithms and Compliance Challenges.” A lawyer on the panel mentioned that a developer wanted to show the lawyer code used for a credit-risk assessment. Because the lawyer could not read code, the lawyer instead requested a memo about what the code did.
This example highlights opportunities for lawyers who know how to read code. It would be much more efficient for a developer to walk the lawyer through the code, at least if it is well written. In addition to the inefficiencies, there is the risk of important information not being communicated when code is summarized in memos. Lawyers with some coding knowledge are poised to differentiate themselves in these matters.
In addition to coding basics, employers value skills in statistics, metrics for legal services, and legal analytics. Many of our LegalRnD students have landed internships and post-graduation jobs because “Quantitative Analysis for Lawyers” was among the classes they took.
Don’t let math phobia repel you from quantitative analysis.
Yes, there will be some math. But law school courses like these usually assume that students have no prior statistical training, which is the way I teach the LegalRnD version of “Quantitative Analysis for Lawyers.” Any law student who has the capacity to craft complex legal arguments can do well in this class.
In addition to learning quantitative basics, students in my class will use statistical tools, such as R, Python, and Tableau, to analyze and present data. All of the material will be presented so that any law student can learn it, if they put forth the effort. For some, this may seem challenging. If so, I urge you to accept the challenge. You will find that the rewards make it worth the effort.
Course Description, LegalRnD version of Quantitative Analysis for Lawyers, Law 637E, 3 credits, Monday & Wednesday 10:30-11:45am
This is an applied course designed to introduce students to various modes of quantitative thinking. The goals of this course are (1) to prepare students to be knowledgeable consumers of quantitative information as practicing lawyers and (2) to prepare students for technology infused law practice of the 21st Century. Course modules include (a) research design, (b) statistics in the courtroom, (c) introduction to probability and basic statistics, (d) data distributions, (e) statistical tests (f) regression analysis, (g) quantitative legal prediction and (h) a brief introduction to legal automation and the technology infused law practice of the present (and not so distant future).