When did public pre-election polling first be introduced (I think Gallup, and a few others were set up in the 30s) ? I've always wondered if the knowledge of likely outcomes disrupts the idea of "voting for the person you like best among the field". And if it does, is that a bad thing ? Consider this scenario : There are 3 candidates A, B, and C. C is a radical newcomer who wants to shake things up. 45% of the people like C. The rest are absolutely afraid of him. A and B are unspectacular guys that share the remaining 55% say as 35% for A and 20% for B. If there were no polling data, and people vote as above, and C wins. If people had access to this information, the B supporters would ditch and vote for A instead, making A the winner. The polls have changed the outcome. So here's the question again : When public opinion polls first started, it was surely known that they would influence the outcome. So was there objection to them; did people think it was a disruption of the democratic process; or was it unanimous that polling was essential to a perfectly democratic process where information dispersal was essential to the integrity of the process ? Alternatively, before scientific polling began, were all elections "imperfect" ? Any information/opinion is welcome.