Mods, are there expiration dates for polling? I would be good to have a good number of respondents for this to be statistically significant, perhaps Morrus can tweet it or something to get a feeling at large what truer numbers are. I'm weary to thinking this data is significant until several hundred have responded, at least.
At a certain point, polling should end though. Shouldn't it? Perhaps not. Stats isn't really my area.
Morrus, what we should really do is ask Mearls et al to have a singular, simple and CLEAR poll on this topic over at their official site on the front page, with public seeing the results for themselves.
That's the only way I'll trust the data, I simply don't trust vague polls that don't ask the real germane questions that are buried away in a survey that takes an hour to do, as being accurate and transparent enough for something so controversial to be settled. (if it even can be, at which point it's time to go back to the drawing board, Mr Thompson)
If I'm in the actual minority, I will sigh and move on, but I want to know. They cannot put such a stinker in the game over the objections of so many of us, without showing us why it had to be this mechanic and not something else in its stead.
If a statistically significant measure that more people like or dislike this mechanic is what you're after, this poll was screwed from the beginning by the inclusion of the "don't care" option. Just a quick check of the standard deviation shows an SD of 12, which means the 7 person difference between "like" and "don't like" is just random noise, unlikely to pass any significance test.
Further, from a experiment design perspective, these polls are the worst kind. They rely on volunteered self-report, which means selection bias is strong. The results cannot be generalized to the larger RPG-playing population. What this poll tells us is that of the members of EN World who have seen the poll, and cared enough about it to vote in it, 7 more people voted for "don't like". Mearls could put up a poll on the WotC site and it'd have the same problems.
For perspective, U.S. presidential election polls generally go for a sample of 1,000 or more. Those polls go through relatively great lengths to obtain a truly random sampling.
WotC's internal surveys are likely much stronger -- Perkins mentioned a 50% participation rate and there have been 150,000 downloads of the playtest, which suggests they've had something around 75,000 respondents. Go ahead and reduce that number to just 10% -- 7,500, and it's still far more than an EN World poll can hope for. Naturally, WotC's surveys still suffer from a certain amount of selection bias -- only those who care to send in surveys are polled.
In as much as 5e goes, I doubt you're in a minority. But I don't think you have a majority, either. At best, one side or the other has a slight plurality, which means that if the designers want to avoid the gnome problem, they've got to split the baby. So damage on a miss is not as widespread as in 4e, it's not a feature all fighters get, like in early versions of the playtest, it just shows up in one ability of a sub-specialty of a one class.
Unless there's a huge backlash against this mechanic that affects overall satisfaction in the final survey, which I doubt will happen, I think it's here to stay. WotC is not polling on each mechanic and then building up the game on the results. They're looking at satisfaction with the game as whole. Even if a plurality suggests they don't like it, but they like the game overall, WotC is likely to just keep it in, on the basis that those who really don't like it (but like the game as a whole) will ignore or houserule it.