Natural Selection

Natural Selection – psychological factors affecting dispute settlement


Psychological studies looking at how people behave when making decisions may seem far removed from the demands of everyday dispute resolution. But, argues James Savory, only by understanding the natural human biases we all have can we support our clients to make the right decisions for them

People in business, including legal advisers, like to think they work in a logical and rational world. But psychological research suggests that we could be misleading ourselves. Sometimes, our attempts to do what is objectively right might get hijacked by unconscious human bias.

Take settling litigated disputes. Of course, it would be wrong to suggest that you cannot get an advantage by hiring a better team than your adversary, but, generally, each side is guided by expert lawyers capable of understanding the law, each side has access to much the same evidence, and each side hires an advocate who can put the case effectively. Judges have long experience in hearing cases and reaching consistent conclusions that are in accordance with the law.

On this basis, in a rational world, you might expect parties to have a clear view of likely outcomes, and to be good at reaching settlements that are consistent with their risks in the litigation. If you were able to look at a comparison between the settlement offers made and the actual outcome of trials, you would expect to find that both sides would turn down an advantageous offer in about the same proportion of cases, and these mistakes would have roughly equivalent financial consequences.

Randall Kiser, in his recently published book Beyond Right and Wrong; The Power of Effective Decision Making for Attorneys and Clients (Springer, 2010), tables his research on whether, in cases where recorded settlement offers were made, plaintiffs (as he calls them) and defendants do make mistakes in about the same proportion of cases, and whether those mistakes have comparable financial costs. Kiser focuses on what he calls “decision errors”. A litigant makes a decision error if the best pre-trial settlement offer available to them was better than the outcome awarded by the court. In such a case, the litigant has missed an opportunity that has cost it a quantifiable amount of money. Of course, in some cases, neither party will have made a decision error. If the court’s award is between the defendant’s and plaintiff’s best offers, they would both, at least in tangible financial terms, be better off by having proceeded to trial.

So what did Kiser find? Across all the cases he researched, the parties made no decision error in only 15%. In 85% of cases, one or other of the parties rejected a settlement offer that was better than the court’s final award.

In a rational world, one would expect the 85% to be spread across plaintiffs and defendants roughly equally. In fact, plaintiffs made decision errors in 60% of cases, whereas defendants made errors in only 25% of cases. Another way of putting it is that, despite rational evaluation and careful consideration of the law and facts, plaintiffs would have been better off tossing a coin to decide whether to accept the defendant’s best offer.

As to amount, while plaintiffs made decision errors in 60% of cases, their errors were, on average, around 8-12% of the final awards. In contrast, where defendants made errors, their average error was in the range of 150-225% of the final awards in all cases studied. So, plaintiffs may get it wrong often, but not by much. Defendants get it wrong less frequently, but when they do so, the error is spectacular. Put another way, the average of the plaintiffs’ best offers was to demand 121% of the final award – not bad, allowing for a touch of optimism. The average of the defendants’ best offers was just a derisory 21% of the ultimate award – suggesting that defendants have not been effectively guided by accurate analysis of their risks.

Kiser considers possible sources of this, as he sees it, bad decision-making and surprising spread of error. He considers some issues more specific to the US, such as the effect of jury trials, and some specific to the lawyers involved, such as the length of experience or law school attended. The possible sources I want to concentrate on in this article are the psychological biases that operate on us, some of which seem to be having a direct effect on our ability to settle disputes.

One thing to make clear here is that the biases emphasised below are all well established. Work demonstrating them has been repeated frequently and is well documented in peer-reviewed academic papers. The interested can follow this up in two articles: Richard Birke and Craig Fox, ‘Psychological Principles in Negotiation Civil Settlements’ (4 Harvard Negotiation Law Review, I, 1999); and Russell Korobkin and Chris Guthrie, ‘Psychological Barriers to Litigation Settlement: an Experimental Approach’ (Michigan Law Review 93, p107, October 1994).

Framing of risky choices

The best way of illustrating this phenomenon is to imagine two groups of similar people – it could be litigation lawyers. One group is offered a choice between receiving £500 or gambling on the toss of a coin; they would get £1,000 on heads and nothing on tails. The other group faces a different choice. Members of the group will either have to pay £500 or they can choose a similar gamble, paying nothing on heads or £1,000 on tails.

If there was no difference between how we view risk as a potential ‘winner’ or ‘loser’, the proportion of each group going for the gamble would be likely to be roughly equal. In fact, in many variations of the same experiment, a far greater proportion of the ‘winners’ elect for certainty, and a far greater proportion of the ‘losers’ take the risk.

In litigation, it is easy to see that a defendant, in reaching a settlement, is being asked to accept the certainty of a loss, as compared to the possibility (however remote) of escaping without loss at all. In contrast, the plaintiff can see the offer of settlement as the opportunity to lock in a certain ‘gain’.

Some suggest that this attitude to risk is based on the relative values we put on things. In a situation where we stand to receive a sum, we value, it is suggested, the £500 difference between receiving nothing and receiving £500 more highly than the £500 difference between receiving £500 and £1,000. In contrast, in the situation where we are facing a loss, it is suggested we value the £500 difference between losing £500 and losing nothing, more highly than the £500 difference between losing £1,000 and £500.

Research goes further. Studies of settlement negotiations show different behaviours on the part of defendants in different types of cases. A comparison was made between contract cases and personal injury ones. Defendants lost 72% to 83% of contract cases, but only 45% of personal injury ones. Surprisingly, defendants were more likely to make no offer of settlement at all in contract cases, where they had a relatively high chance of losing, than in personal injury cases. This suggests that defendants are actually more likely to take a risk on trial where their chances are worse!

Attribution error

Humans are storytellers. We look for meaning and explanation behind the events that affect us, and then base our reaction on that meaning. And the stories we tell can be shown to vary, depending on the circumstances. Take, for instance, cases where what we have done has upset or hurt someone else, compared to where what they have done has upset or hurt us.

On the whole, where we have done something to hurt another, we look to explanations outside ourselves. We were forced to do that because of the way someone else behaved towards us. The road was unexpectedly wet or icy, rather than we were driving too fast. In contrast, we see the explanations behind what other people do to us as being attributable to their personal characteristics. The other driver involved was reckless, paying no regard to the safety of others. The business that failed to deliver our goods on time knew they would not be able to do so, even when we placed the order, so they defrauded us. And, of course, the fact that they disagree with us just demonstrates that they are untrustworthy. The emotional effect this attribution has on the way we conduct ourselves in litigation is dramatic.

Another example is the attributions we make where we win litigation as compared to where we lose.  Where we win it is because we made good decisions, interpreted the evidence correctly and argued the case convincingly.  Where we lose, the judge failed to understand the case, was friends with their advocate, and believed the lies their witnesses told. And the danger of these stories is that we fail to learn from setbacks and reinforce over-optimism as to our skills.

Indeed, over-optimism is almost a category of bias in its own right. What proportion of you would say you are a better than average driver? Studies predict that 90% of you will think so. And you are in good company; in one study, 90% of judges estimated that “at least half of their peers had higher reversal rates on appeal”.

Reactive devaluation

It seems clear that our reaction to a proposal is coloured by who it comes from. In the classic experiment of the 1980s, Americans were asked to evaluate a nuclear disarmament plan. One third of the group was told the plan came from President Reagan; another third that it came from independent neutrals; and the last third that it had come from President Gorbachev. 90% of those that were told the plan had come from Reagan regarded it as satisfactory. 80% of those who were told it had come from neutrals felt the same. But only 44% of those who were told it had come from Gorbachev did so.

In settlement negotiations, our mistrust might be linked to the attribution of motives mentioned above. For instance, if we see the other side as malicious, rather than the victim of circumstances,when they put forward a reasonable proposal, we may imagine they know something we do not.

Another explanation is that, if we have become deeply mistrustful, we may begin to give value to denying them what they want. If they advance a proposal they will be happy if we accept. But we do not want them to get what they want, so we reject it, even though, before it was made, we might have regarded it as satisfactory. This certainly makes interest-based negotiation more difficult.

And there is some evidence that, in any situation, a solution offered becomes valued less highly simply because it is offered. So, in role-playing experiments where a control group was relatively indifferent between two possible settlements of a dispute, groups who were told that one or other of them had been offered veered towards preferring the one not offered. Another way of putting it is that we value more what we cannot have than what we can.

Status quo bias

It also seems, slightly inconsistently, that we value more highly something that we already have, as compared to something we might acquire. Again, this bias has been demonstrated many times in tests. The classic example is the ‘mug and chocolate experiment’.

The group was divided into three. One third of the group was given a choice between being given a mug or a bar of chocolate. 56% chose the mug and 44% the chocolate. The other two groups were allocated the mug or the chocolate, and members of those groups were then asked whether they would like to swap the mug for chocolate or vice versa.

Based on the choices made by the control group, you might expect roughly half of those who had been allocated one of the items to ask to swap to the other. In fact, the proportion wanting to swap was only 10 or 11%.

So, in litigation, we are less likely to give up what we feel we already have, even if we are offered more to do so than we would have been prepared to pay to get it.


We all know that plaintiffs will make high demands, and defendants will seek to counter-claim, in order, in each case, to shift the range of bargaining in their preferred direction. What may not be so well understood is just how powerful the psychological effect of so-called anchoring really is.

The surprising truth is that people’s behaviour is affected by a number that they have been exposed to, whether that number bears any relationship to the context of the behaviour or not. This has been demonstrated in many cases.

One example is of two groups who were exposed to a spinning wheel that was rigged either to stop at the number 10 or the number 65. They were then asked what proportion of countries within the UN came from Africa. On average, those exposed to the number 65 estimated 45%. Those exposed to the number 10 estimated 25%.

In the context of settling litigation, it barely seems necessary to mention research in which the relative happiness of plaintiffs faced with an offer from defendants of $12,000 was measured. The group that had previously been offered $2,000 rated the new offer far more highly than the group that had previously been offered $10,000.

Selection error

Things are happening around us all the time. Our senses are bombarded with information. If we tried to pay attention to everything, we would be overloaded. So, our brains are trained to choose. Most of the time, the system works very well. When we are driving, we notice things that might be a threat – the child playing with a ball, the bonnet of the car emerging from the side road. But the way in which our brain chooses can also lead us astray.

We discover that we see or notice what we expect, or want, to see. So, for instance, supporters of competing college football teams, asked to review a video of the game, invariably saw the other side as having committed more fouls.

So what we see, remember and notice begins to get moulded by what we want to see and would like to believe to be true. This seems to mean that the longer we investigate and build our case, the more faith we create in the truth of our story, and a belief that there can be no evidence to overturn it. As we unearth new evidence, we notice and give weight to the elements that support our case, and fail to notice or draw out the elements that lean in the other direction, thus making it less likely that we will find a settlement point.


In this article, I have drawn attention to some of our inbuilt psychological biases that seem to have a significant effect on the way we try to settle litigation. We are not powerless in the face of these pressures, but litigants and their advisers do need to make special effort to avoid them having an impact that is later regretted. Mediators who are familiar with these human tendencies also need to design their processes and interventions to minimise their effects.