If the hypothesis passes the significance test then it is treated as being true, but a lot can go wrong.

FORBES: Strategy at Digital Speed

Only one of those is moving in the right direction above the level of the test of statistical significance.

FORBES: Jeff Sach's Lancet Paper on Millennium Villages: Actual Scientists Seriously Unimpressed

The bad news: The new Genentech test did not reach statistical significance, meaning that it does not pass a mathematical litmus test to prove beyond a doubt that the result was not based on chance.

FORBES: Genentech's 'Bad' Data Actually Good