Previous Entry Share Next Entry
Poll-Taking Behavior
Poll #1362268 Poll-Taking Behavior

What do you do when none of the provided answers to a question are quite right?

Select the closest answer.
Select two or more answers that bracket the real answer (if possible).
Select the answer that gets you the result you want. ("None of these are what I'd do, so I'll pick the Ravenclaw / Captain Jack Sparrow / INTJ thing to do.")
Skip the question.
Contact the poll creator.

What do you do when a poll does not allow you to skip a question you'd like to skip?

Select a random answer.
Quit taking the poll.
Contact the poll creator.
I don't skip questions.

What do you do when a poll does not allow you to select multiple equally-correct answers to a question?

Select one at random.
Quit taking the poll.
Contact the poll creator.
I can always find one best answer.

When you take a poll with answers in the form "1. strongly agree 2. somewhat agree 3. neither agree nor disagree 4. somewhat disagree 5. strongly disagree", which numbers do you usually end up picking?

Mostly 1s and 5s.
Mostly 2s and 4s.
Mostly 3s.
Equally distributed.
I haven't noticed a pattern.

  • 1
To elaborate on the answer to the first question, I'll select a reasonably close answer if there is one (which is most of the time), or otherwise skip it. On the second question, there wasn't a reasonably close one -- I usually pick the response that seems least likely to skew the data (which is sometimes skipping it, sometimes picking the "meaningless" answer).

And, yes, the fact that I answered "select the closest" to the first question, and then proceeded to not do that on the second, did strike me as ironic. But it was clearly the most accurate set of responses, so....

I considered deliberately setting up the responses so there wasn't a consistent way to answer them. But then I figured it'd probably happen naturally anyway.

1. If it's an LJ poll, I give up on trying to fill it out and just post a comment instead. :-)

For all of the first several questions, really, it depends a lot on the poll. (I assume we're talking here about online "Which Jedi Knight Are You Most Compatible With?" kinds of things, not surveys encountered in other sorts of contexts.) If it's generally an interesting, good-quality questionnaire that just happens not to fit me in places, and the problems aren't serious enough to make the results pointless, I make do with the options available; if I really can't convince myself that one is more apt than the others, I pick pretty randomly. If it doesn't seem worth the hassle, I quit. I've never contacted a poll creator.

On your last question, I think I do gravitate toward particular answers, but which ones depends on my mood and on what I'm filling out. Sometimes it's one of the three options you give; sometimes I'm disposed to be more emphatic about agreement than disagreement, or vice versa, and so it comes out more "1s and 4s" or "2s and 5s."

I generally only contact poll creators if it is some sort of survey that's actually looking for information, because I figure they want to know they left out something important. (And I usually only do that if they have a space for comments in the poll somewhere.)

sometimes I'm disposed to be more emphatic about agreement than disagreement, or vice versa, and so it comes out more "1s and 4s" or "2s and 5s."

Interesting. The ones I see that use that format often have a substantial fraction of the questions phrased negatively. I always found that confusing, but perhaps it's to adjust for this kind of tendency.

(Deleted comment)
I like including "Other" or "N/A" and having some room for folks to expand on that. It's hard to think of a comprehensive answer set!

LJ polls are easier since I can just pick the closest answer and then explain what I /really/ mean in the comments. Polls on other sites are much more likely to be quit.

i have very low patience with polls at the best of times, so if the options look wrong to me i just bail out.

Some of these need to be checkboxes rather than radio buttons. For example, what about "select the closest answer but also contact the poll creator"?

Hm. Well, you've done part 2....

Indeed! :) Whether part one happens or not depends on how ornery I am. :)

Man, this poll really toyed with me.

The answer to the first question that I wanted to give but couldn't was "Quit taking the poll". Which I almost did, but then I just decided to skip it instead so that I could say "Quit taking the poll" for questions 2 and 3. Then I got to question 4, and I almost quit the poll again, since my answer wasn't there, but instead I submitted the poll and came to bitch in the comments.

My answer to 4 is "I try to answer accurately, but I consider how the poll is going to be used and bias my answers accordingly. For example if the poll is to be used to decide whether or not to improve service, I'll vote that the service can be improved; my choice being at that point between "It was terrible" (1) and "it was ok but could be better" (3). If the poll is to be used to decide whether someone should get a pay raise, though, I'll usually bias the other way, deciding between "They are doing an ok job" (3) or "They are fantastic" (5). If the information isn't one that I usefully decide how to bias, then I reply across the whole range accurately.

So for the question about what you do when there isn't anything that's quite right... there wasn't anything that was quite right... so you skipped the question... rather than selecting "skip the question". Ow, my brain.

This reminds me of the self-referential multiple-choice logic puzzle.

That's an interesting point about deliberate bias based on the use of the poll.

Depends on the purpose of the poll. If it's something I need to complete for some reason I care about, I either find the best answer, check two answers, or contact the test author. If it's something I was doing for fun, I just stop taking the poll.

  • 1

Log in

No account? Create an account