Statistical evidence is admissible to establish predominance under FRCivP 23(b)(3) if that evidence would be admissible in an individual action on the same claim, the statistical evidence is linked to the plaintiffs’ theory of liability and the use of averaging assumptions does not conceal the variations that otherwise would defeat class certification. Here, plaintiffs’ statistical evidence satisfied those three tests. statistical evidence is admissible in individual antitrust price-fixing suits to show the individual plaintiff’s damages. The statistical evidence matched the price-fixing theory of liability, and the averaging assumptions did not conceal disqualifying variations, but merely differences in the amount of damages sustained. However, on class certification, the plaintiff bears the burden of proving by a preponderance of the evidence that common issues of law and fact predominate. Here, the district court simply decided that the plaintiff’s statistical evidence was admissible, but didn’t determine whether it satisfied plaintiffs’ burden of proof. In particular, defendants produced evidence showing that the plaintiffs’ statistics showed that 28% of the proposed class suffered no injury. If true, that fact would defeat class certification since a class may be certified only when a de minimis percent (i.e., less than about 6%) of class members are not entitled to recovery. The district court had to resolve this factual question before certifying a class.