Dateline: November, 2021, Issue 2

How accurately do jurors decide liability based on preponderance of the evidence?

In most civil litigation, the burden of proof required to find a defendant liable is probabilistic: If it is more likely than not (i.e., over 50%) that a defendant caused a plaintiff's injury, the defendant is to be held liable for that injury.

Lariviere (2015) investigated the willingness of mock jurors to assign liability if the likelhood the defendant caused a plaintiff's injury was either 5%, 50%, 51% or 95%. Mock jurors read written case vignettes about a plaintiff company that purchased ball bearings from the defendant company for use in 100 rocket motor tests. The number of bearings supplied by the defendant company varied from 5 to 50 to 51 to 95, with the remaining bearings coming from the plaintiff's stock purchased from other equally reputable manufacturers making identical guarantees about the quality of their bearings. In all cases, mock jurors read that one bearing failed and that it was unknown whether the bearing came from the defendant or from the plaintiff's stock. Immediately before making their liability decision, mock jurors were told that the standard of proof in civil litigation is that if it is more likely than not a defendant has caused damage, then the defendant is liable for the harm and damage caused. Mock jurors were also asked for the reasons for their liability decisions.

More jurors held the defendant liable when the evidence was sufficient than when it was not sufficient, although only 57% of mock jurors correctly assigned liability, that is (a) finding the defendant liable when the evidence against the defendant was sufficient (the vignette likelihood was 51% or 95%) and (b) not finding liability when the evidence against the defendant was insufficient (the vignette likelihood was 5% or 50%).

Many jurors judged the defendant liable when the evidence was insufficient to do so. When the likelihood the defendant caused the plaintiff's injury was 5% or 50%, a significant percentage of mock jurors given those vignettes incorrectly held the defendant liable (45% and 54% respectively).

Additionally, many jurors judged the defendant not liable when the evidence was sufficient to do so. When the likelihood the defendant caused the plaintiff's injury was 51% or 95%, a significant percentage of jurors given those vignettes incorrectly decided the defendant was not liable (39% and 35% respectively).

Every mock juror was instructed to use the provided civil burden of proof standard to assign liability; only 25% reported doing so and fully 96% of them made correct liability decisions. By contrast, 25% of mock jurors reported using the criminal standard of beyond a reasonable doubt to reach their liability decision. Many mock jurors reported wanting more or different evidence before assigning liability and only 45% of these mock jurors made a correct decision as to liability. The willingness of mock jurors to use the preponderance of evidence standard they were provided -- not ignorance, not lack of care, not lack of understanding -- was found to be the most important factor in predicting correct liability decisions.

McKinlay (2017) conducted a similar experiment with law students serving as mock jurors. All mock jurors were told that two defendants (Company A, Company B) supplied ball bearings for the Plaintiff's rocket motor, that the rocket motor exploded, and that the telemetry system identified that the cause of the explosion was solely attributable to the failure of a single ball bearing, although the specific supplier of the failed bearing could not be determined by any post-explosion investigation. Three variations of the case vignette were created based on how many bearings came from Company A versus Company B: 70%, 51% or 50.01%. Mock jurors were told to use the preponderance of evidence as the standard of proof and were provided its definition. All mock jurors were required to recognize the correct definition of preponderance of the evidence prior to making a liability choice from among one of 5 provided options: (1) neither company is liable, (2) there is insufficient information to make a decision, (3) both companies are 50% responsible, (4) both companies are liable corresponding to the percentage of bearings supplied by each company, or (5) Company A is liable.

Despite sufficient information being provided to assign liability, 22% of mock jurors in McKinlay's research indicated they needed more information before liability could be decided. Despite Company A being the only correct choice for liability (as only Company A supplied the preponderance of the ball bearings in all vignettes), only 24% of mock jurors correctly assigned liability to Company A. The most popular liability decision was to assign liability based on the number of bearings supplied by each company, which was selected by 34% of the mock jurors, which McKinlay notes seems intuitively fairer though is legally incorrect. McKinlay suggests that "jurors must be persuaded that the preponderance of evidence standard is fair, and while it does not necessarily reflect the 'truth', it is just and worthy of application" (p. 37).

Both McKinlay and Lariviere report that other studies have found similar results using different methods and different case vignettes: that is, too few people assign liability correctly based on the standard of preponderance of the evidence.

In sum, liability decisions are fraught with inaccuracies. Defendants are incorrectly held liable and incorrectly not held liable. Jurors use incorrect burdens of proof despite knowing the correct standard. Intutions of fairness compete with legal burdens of proof.

Source Lariviere, K.L. (2015). Mock jurors made mistakes assigning liability even though the civil standard of proof and the evidence were clear and precise: Mock jurors set aside the standard of proof. Unpublished Master's thesis. Saint Mary's University, Halifax.

Source McKinlay, J.A. (2017). Adding a numerical description to civil standard of proof jury instructions: Probabilistic evidence still defies correct liability assignment. Unpublished Master's thesis. Saint Mary's University, Halifax.