Try as they might, researchers aren't having much luck finding ways to teach lay jurors (or judges, or lawyers) how to draw sound causal judgments from scientific studies. Our causal heuristics have failed to keep up as we've moved beyond discerning why ripened apples invariably fall (rather than float away, etc.), to probabilistic judgments made under conditions of deep uncertainty - judgments hinging upon the unsettling notion that only some people, in some circumstances, experiencing some exposure, to some thing, have a somewhat increased risk of something else.
Short courses designed to bring lay jurors up to speed on sound causal reasoning managed, in one iteration of the study reported, to make the jurors slightly more pro-plaintiff (when the scientific study in question reliably pointed to a pro-plaintiff outcome). But the would-be jurors in other scenarios remained blind e.g. to the need to rule out an obvious alternative causal explanation. Specifically, despite the fact that the jurors were told that a bromine compound to which plaintiff was also exposed was itself perfectly capable of causing plaintiff's illness, the jurors failed to demand that it be "ruled out" before settling on the putative cause (beryllium) as the actual cause. "[E]ven trained jurors overlooked an obvious confound in which participants in an experimental group were simultaneously exposed to two different chemicals, with the resulting detrimental effects attributed to only one of the chemicals (defendant's)."
In the end, the mere ipse dixit of a credentialed expert willing to say that a hopelessly flawed study established causation was, over and over, enough to carry the day for plaintiff.
Ultimately, counseling mock jurors that cause ought to precede effect and that effects ought to be correlated with putative causes before causation is imputed improved their reasoning; but only a little. It was still the case that even when plaintiff had no (zip, zero, nada) credible evidence of causation a juror was inclined to rule in plaintiff's favor more than a third of the time. Any plausible scientific study tending to support the hypothesis of general causation was enough to move the needle past 40% before specific causation was (apparently) even addressed.
So again we see why judicial gatekeepers will remain indispensable to toxic tort litigation. We could subject jurors to a short course on statistical inference and nudge them, perhaps, a bit closer to sound causal judgments but we'd still end up giving one side an enormous advantage. Have a look at the tables starting on page 31 of the study if you have any doubt.