ODR: A Time for Celebration and Procedural Safeguards

Nancy A. Welsh, Professor of Law and William Trickett Faculty Scholar, Penn State University, Dickinson School of Law, and Chair-Elect, ABA Section of Dispute Resolution

There is no question in my mind that those of us involved in ADR need to understand and embrace online dispute resolution (ODR) and celebrate its potential. Meanwhile, I urge ODR’s advocates and implementers to learn from the experience of ADR’s advocate and implementers, and to be wiser than we were. Specifically, I applaud ODR advocates’ passion, innovation, and optimism, and I hope that you will see procedural safeguards as helpful in achieving your goals, not as unwelcome signals of distrust.

To that end, I will focus on a concept called “procedural justice”—i.e., what it is, why it matters, to whom it matters, whether it assures fair solutions, and what it suggests for best practices in ODR. I will also raise issues regarding the underbelly of procedural justice—i.e., how it can be manipulated, by whom, why and when that matters, and what it counsels regarding procedural safeguards. My goal is to suggest how procedural safeguards may help with avoiding or, at the very least, quickly addressing worst practices in ODR.

So, to begin, what is procedural justice? First, it involves people’s perceptions regarding whether a procedure is fairnot whether it satisfies them or makes them happy. Second, it is a socio-psychological concept, with a vast empirical literature detailing its application to the courts, mediation, negotiation, interactions between people and police, and in the workplace. Third, this empirical literature reliably reveals that people are more likely to perceive a process as fair if it provides for the following:

  • Voice – the perception that you were able to express what was important to you.
  • Trustworthy consideration – the perception that the authority figure (or the other decision-maker in a negotiation) cared enough to listen and tried to understand what you had to say.
  • Neutral forum and even-handed treatment – the perception that the forum is neutral and that you and your claims were treated in an even-handed way.
  • Dignified treatment the perception that you were treated in a dignified and respectful manner.

Research shows that if a procedure includes these characteristics, most people are likely to perceive that the substantive (or “distributive”) result is fair, even if it is adverse to them; they are more likely to comply with that outcome; and they are more likely to perceive the institution hosting the process as legitimate—which obviously is important to a public institution like the courts.
Do fair procedures assure fair outcomes? It does seem logical that if parties are permitted to have voice about what matters to them, and if they receive trustworthy consideration, and if the forum is neutral, even-handed and dignified, the outcome should be better-informed. This logic also represents a heuristic, of course, a mental shortcut used to assess distributive fairness. Similarly, it seems that if parties are permitted to have voice, if they receive trustworthy consideration, and if the forum is neutral, even-handed and dignified, then they will be treated as “in-group” members – and that generally means they will be treated fairly. Again, this seems logical.
Do we know, though, that fair procedures actually assure fair outcomes? Of course not. In some sense, equating fair procedures with fair outcomes represents an act of faith. Further, we must admit that we cannot always know what represents “the” fair outcome. There may be more than one fair outcome, depending upon the norms used.

That brings me back to the relative consistency of procedural justice and what it suggests regarding concrete questions and considerations to achieve best practices in ODR:


  • Does the ODR process provide people with the opportunity to express themselves? Can they input information, particularly information that matters to them? As they respond to questions, will they be limited to a pre-determined list of choices or will they be able to choose “other” and input customized information? Even better, will they be asked open-ended questions?
  • Is the medium (text-based, audio-based, video) one that is comfortable and accessible for everyone? I know people who do not know how to use computers and do not own smartphones. I know people who really need to talk with a human being when a problem arises. If these people are not comfortable with ODR, will they have reasonable access to the “old fashioned” hearing that they prefer? Will they ever be able to be in contact with an actual human being? Even though I support ODR, I think it is absolutely necessary that government offers multiple “channels” or “paths” to justice. I worry about systems that reserve “old fashioned” human contact for the cases that involve a sufficiently large amount of money.

Trustworthy consideration

  • Does the ODR process demonstrate that it has “heard” what people had to say? Do the questions and information flow follow logically from what people expressed? Does the screen display a summary of what someone input? Does the process provide an opportunity to correct that summary? Does the process provide an opportunity to observe an actual human being’s reaction?
  • If the ODR process produces a decision or award, does the award include an explanation regarding why the ODR process is providing this result? Does the explanation include some reference to the facts or other information that the person input?
  • If emotions were solicited and expressed, can we trust that they were heard and understood? And is such understanding being expressed by an actual human being or is an algorithm or Artificial Intelligence providing a proposed or final response? If it is the latter, I am uneasy describing the dynamic as “trustworthy consideration” unless the ODR provider acknowledges the use of algorithms or Artificial Intelligence, and continues to offer a means to reach an actual human being.

Neutral forum and even-handed treatment of parties and their claims

  • Has the ODR forum been developed by a third party, or by one of the parties to the dispute? The former is more likely to be perceived as neutral. The latter is more likely to be perceived as biased and thus subject to distrust.
  • Is there transparency regarding how the process works and produces individual outcomes, as well as transparency regarding its aggregated outcomes?
  • Are the parties able to “see” the ODR process dealing with both of them at the same time—and be assured of its even-handedness?

Dignified treatment

  • Does the ODR process permit people sufficient time to input information and reflect on their responses? Does the process permit people to change their responses before they ultimately click “submit?”
  • Does the ODR interface convey a sense of both accessibility and dignity to people participating in the ODR process?  Does it convey a sense of dignity to people observing the process?

Now we will turn to the underbelly of procedural justice.  Simply put, procedural justice can be manipulated.  Let’s assume I am the decision-maker in a procedure. I can give someone the opportunity to speak; I can sit and look like I’m listening; I can promise that I will consider what this person had to say; I can be polite. But it can all be an act. I may have no intention of allowing this person to influence my decision-making. Or I may have input certain information into an algorithm beforehand and have no plan to deviate from the algorithm’s recommendation. The fix is in.

People know they can be manipulated, and ODR must contend with the heightened suspicions caused by others’ misbehavior in the worlds of Big Data and social media.  So while I urge ODR advocates and implementers to consider the questions and issues listed above, I also urge you to go further and embrace procedural safeguards—e.g., disclosure of how your processes work and what norms are embedded in your algorithms, transparency regarding individual outcomes and aggregated outcomes (and as differentiated by salient demographics), willingness to develop something akin to algorithmic “audits.”

Owen Fiss famously declared himself to be “against settlement.” At the time, I was offended. How could he have so little regard for the value of ADR? Today, I wonder whether I should have listened–not in order to be dissuaded from pursuing ADR–but in order to try to anticipate and prevent likely problems and abuses.

Today, I see and celebrate the promise of ODR. And as a friend, I plead with ODR’s advocates, implementers and providers to be ready and willing to anticipate the worst. Be ready and willing to anticipate that your good processes, with all of their promise, could also be used for ill. Western literature, old and new, is replete with tales of fallen angels.  So embrace those procedural safeguards.  They may be pedestrian, but they are also essential to achieving the exciting potential that your pioneers and prophets have foretold.


This is a shortened and revised version of a contribution to the 15th ODR conference held in The Hague in May 2016.

Leave a Reply