Received from Duane at 10/3/2018 04:02 PM UTC:
On Wed, Oct 3, 2018 at 10:56 AM, Jim Higgins wrote:
This is something that Y! did right.
Not really. The only time I edit a poll is if I think of an additional answer that may be selected by a lot of people. The existing choices don't change, so aren't completely void, though may not be completely accurate either.
But accuracy matters!
It's easy enough to let folks know that the options have changed, so they can change their answer if they want to. I haven't tried changing existing choices, so I don't know what would happen with your example.
Duane
There's a change notice sent. But what percentage of those who have already voted will read the change notice? Not enough based on my experience with communicating anything via group email.
After receipt of your reply, I created a poll in one of my groups.
Initial poll: Which do you like better? 1) Chocolate Cake, 2) Brussels Sprouts
I and one other member voted within a couple of minutes. Two votes for chocolate cake.
Then I thought of "an additional answer that may be selected by a lot of people," so I edited the poll...
Edited poll: Which do you like better? 1) Chocolate Cake, 2) Brussels Sprouts, 3) My child
I changed my vote to "My child."
I waited about an hour... and then closed the poll.
Results? One vote for My Child, one for chocolate cake.
Had I waited a couple of days before changing the poll, I'm fairly sure the one other guy who voted - someone I know pays attention - would have either changed his vote or phoned me to ask "What the heck?" Many others who pay far less attention than he does would have voted in the initial poll for cake or sprouts and many of those wouldn't have noticed when the poll changed, or would see the title on the change message and assume it was just a reminder so they wouldn't even read it much less change their votes. So had I taken votes for several days before changing and then waited several more before closing there's absolutely no doubt in my mind that there would be many votes for Chocolate Cake and not so many for My child.
But never mind... if I run a poll and need to change something midstream I can delete the first poll and start a new one. Good luck to the pollsters who don't understand that midstream changes make for inaccurate polls. I mean, why run one if it's going to be run in a manner that assures inaccuracy? Not to mention the inaccuracies that are outside our (and Gio's) control, such as self-selected responders.
Jim H
