(3rd Cir., filed October 18, 2016): Arguing that an expert cannot premise a causation analysis on a single statistically-significant association when the larger body of epidemiological studies fails to find any such association. Also, arguing that an expert cannot massage the data with after-the-fact analyses to create associations that were not found by the statistical methodologies originally selected by the scientists who performed the study. Trial court judges must act as gatekeepers over the reliability of expert testimony, carefully evaluating whether such testimony is based on sound scientific principles or is simply bought-and-paid for “junk science.”
