There is no moderately easy way to test this hypothesis scientifically as you have to control for all other significant variables. 1) You don't know what these variables are 2) You have no way to control for them.

@AaronKreider That said, unless the margin of popular vote victory is 0.5% or less - this should resolve No.

I've update the ** to require larger increases for larger prior probabilities and lower ones for lower probabilities. I think this makes sense because if the candidate starts at 20% and Trump's endorsement gives them a boost to 25%, then his endorsement constitutes a fifth of their chance of winning. However, for 60% -> 65%, Trump only constitutes 1/13 of their chance of winning. It doesn't make sense to require a higher burden in the lower probabilities when the candidates win was made much more likely by Trump's endorsement. And likewise it doesn't make sense to require a very low increase in the high probabilities, when the candidates win was already very likely.

To offset the negative effect for NO shareholders of this looser criteria, I've set a limit order for NO at 20%.

The ** does a lot of work but if a canidate was already at 70% then can we really say they won because of an endorsement? At 70% it would be relatively safe to say they could have won on their own. Winning more is not the same as winning because.

@ShitakiIntaki True. In retrospect I should have said the opposite. Smaller increases for lower probabilities, and larger increase for larger probabilities. It might not be too late to change it.

@xyz just relax the standard to "Will you (the market creator) believe that Trumps endorsement helped a canidate enough to get them the win" and then keep the examples you have of what you would believe based on the various scenarios.

In some senarios you might observe a drop in polling numbers after an endorsement in which case you might believe the endorsement hurt rather than helped the canidate and then if they canidate goes on to win anyways you might believe they won inspite of a damaging endorsement, or information may come out later that leads you to believe it helped in the long run.

Causality is a high bar to "prove" ultimately you are more likely to be left with sentiment rather than proof.