« JELS, 3:1 | Main | New Book - The Politics of Precedent »

01 March 2006



The issue of lawsuit quality is of great empirical interest. I believe cognitive psychology would have a great contribution to make not only in regards to understanding jury decisions, but also pretrial decisions to settle.

My view is this: Hindsight bias (HB) theory predicts that the more extreme the outcome, let's says profound cerebral palsy, the greater the HB effect. Jurors, or attorneys, would be more confident in "just knowing" that human error (read "negligence") was responsible in such cases than milder CP cases, given similar ambiguous antecdents.

I believe attorney's like Feiger are successful precisely because they take advantage of HB by choosing the most severe outcomes regardless of antecedents.

The reason I came to this blog was to ask whether there is any empirical research looking at severity of physical outcome as a predictor of legal outcome. Can anybody point me in the right direction, or alternatively stop me from wasting my time looking for something that has not been done?

frank cross

It's a good question that should make us all cautious about interpreting our research. I think it may infect some of the findings of the research on religious freedom outcomes being discussed over on Volokh.

I think there are two main considerations. First, is there some good theoretical reason to expect cases to be of differential strength (e.g, the civil/criminal distinction above). Second, is there a general economic selection effect theory. E.g. Priest/Klein suggest that only cases at the margin are litigated to a decision, but we know this is not exactly true for several reasons such as asymmetric stakes and asymmetric info.


I was merely noting civil/criminal differences as a non-controversial example where one would expect different rates of success for different cases ex ante.

For med-mal, my model would predict that the injured-baby/cerebral-palsy cases in a state without caps are more likely to be weak cases, because a lower p is needed to make the expected value of suit positive. One need merely look at the track record of a Fieger, who clearly treats his med-mal cases as a portfolio--someone who was confident in the merits of his case wouldn't do so much to risk reversal.

Unfortunately, one can't entirely judge the merits of a case by virtue of settlement or jury decisions. There is the complication of outcome bias by jurors, which in turn affects settlement decisions; there's also the complicating factor of post-trial bad-faith-refusal-to-settle litigation judged in hindsight for an insurer that refuses to settle at policy limits. Through backwards induction of the game-decision-tree, this also changes the likelihoods of bringing a case.

(I'm thus amazed that someone as sophisticated as Vidmar would suggest (as he does in his most recent paper) that reform that changes trial rules won't have much effect on medical malpractice litigation because most cases are settled. I hope that was just a poorly phrased conclusion, rather than something else.)

There's also the problem that any method that attempts to qualitatively evaluate the p-value of an individual case invites the question why we aren't using that qualitative method instead of a jury trial.

Jeff Yates

This is a really good question and brings to mind a number of related threads of research that have attempted to, in one way or another, control for quality of claims, either at the trial or appellate levels (although typically the latter).

For starters, Galanter's classic L&SR piece notes that the claims of have nots are typically inferior to the claims of haves, in part because haves are typically repeat players and thus a) bring more experience to the game, and b) are playing for the long term, and therefore settle their weaker cases so as to not set bad precedent, etc. (I'll admit that there might be long term rationales for litigating in weak cases as well, e.g. discourage future plaintiffs).

At the appellate levels a number of tactics have been used to assess, either directly or (more likely) indirectly, quality of claims. McGuire (1998)and others have looked into attorney experience as an indicator of, well, the impact of good legal counsel (which may improve quality of claim). Also, there may be an aspect of lawyer selection effects as well. He showed that the SG wasnt really any more successful than other litigants after you controlled for experience.

Another method is to model factual cues and legal claims (or subclaims) that tend to be more successful than others within a certain ligitation context (e.g. certain affirmative defenses, certain factual contexts, etc. - such as Segal 1984 APSR; also McGuire on obscenity cases and Songer at Ct App level).

Of course, party type has received a good deal of attention. Sheehan and Songer modeled it at the US Ct App level and the SCOTUS level, finding that have nots do poorly at the ct app but not as poorly at the sct level. Epstein and Rowland have looked at the impact of interest groups as litigant types at the US Dist Ct level, using matched pair methodology.

On Michael's specific question - how would party type matter in med mal - I'm not an expert on med mal, but I would guess that there might not be a great deal of variance among plaintiffs - most are individuals and one-shotter litigants; most defendants are likely similar I suspect. You could do like McGuire and look at lawyer experience or specialty. I'd imagine that this might make a difference. Also, certain legal factual cues or contexts might make for a more quality (i.e. success probability) claim. So, perhaps one type of med mal injury is a "winner case" and another rarely succeeds - either due to standard of care issues or, perhaps, typical jury sympathies.

Just some thoughts on this matter -- On a related concept, McGuire and Stimson (2004) make an interesting case that in assessing SCOTUS liberalism we should only consider the sct's reversals because it's a better indicator of preference than aggregated cases or affirmances - which represent error on the part of the appellants.

Michael Heise

Good point, and I am inclined to agree with you about expected differences across criminal and civil sectors. What prompted my question, however, was a more narrow context (e.g., whether party type differences will express in terms of lawsuit quality in the med mal context--that is, wholly within the civil sector and where the legal rules are held constant).

Ted Frank

Put slightly differently, is there any reason to assume, ex ante, that lawsuit quality will systematically vary across party types?

Yes. To take an obvious example, criminal appeals are subsidized by the government, while civil appeals require a civil party to devote its own resources. So, ex ante, one would expect civil appeals to have a higher success rate than criminal appeals, because low-value civil appeals are priced out of the market and not made, while a long-shot criminal appeal will usually have a value to the party greater than the zero cost that the criminal defendant incurs. And, indeed, one sees that result, with civil appeals having higher success rates than criminal appeals.

One can construct game-theory models for different types of litigation and quickly see that changes in legal rules (loser-pays, damages caps, early-offer settlement requirements, heightened pleading requirements) will have a substantial effect at the margin of whether to bring a case with a low chance of success.

The comments to this entry are closed.


May 2018

Sun Mon Tue Wed Thu Fri Sat
    1 2 3 4 5
6 7 8 9 10 11 12
13 14 15 16 17 18 19
20 21 22 23 24 25 26
27 28 29 30 31    

Site Meter

Creative Commons License

  • Creative Commons License
Blog powered by Typepad