A lot of the contemporary formal scientific process is done incredibly badly, for a variety of reasons including overt political bias on the part of individual scientists working in the academic system, pressure to publish any results including poor ones, and outright laziness and fraud. In general we shouldn't assume that if a bunch of public scientific studies purport to show that some phenomenon is happening, that that phenomenon is actually happening. It takes substantial time, effort, and experience to evaluate whether a claimed scientific result is valid; and all the moreso when that result has immediate political policy implications.