Why? This wasn’t a fishing expedition where they took two cohorts of marines and looked down a list of 20 diseases to find one with p<0.05 distribution between

Why? This wasn’t a fishing expedition where they took two cohorts of marines and looked down a list of 20 diseases to find one with p<0.05 distribution between the two cohorts. > Occupational exposure to the industrial solvent trichloroethylene (TCE) was previously associated with a 6-fold increased risk of Parkinson disease (PD) in a small study of twin pairs discordant for PD.1 Animal studies provide biological plausibility for this association by recapitulating key pathologic characteristics of PD Given this - a previous finding, a plausible mechanism - they went and found two good existing cohorts who had different exposure to the compound of interest and for whom they could access medical records. That is good study design. Which is probably why it got published.