I happen to spend a bit of time doing online surveys, whether they’re for website feedback or market research or, well, anything else.
Whilst I haven’t done any in depth research I’ve certainly noticed a lot of basic mistakes that are really annoying! So here’s a few that you should avoid and you should make your surveys easier for people to fill in and that in turn should improve the quality of your results.

1. Don’t be too repetitive

Sure you’ll need to repeat some things just to cross check them and a lot of the time surveys make use of subtle repetition to check people are giving honest answers. But if people are faced with the same questions over and over again, they get bored, they don’t read the questions properly and they start skimming and just clicking through to get to the end of the survey and some people leave altogether.

As an example I was filling in a market research survey about drinking habits and, well, I drink a lot of tea – seriously – a lot! So in this survey it had noted that I had drank about 9 cups of tea in the past 24 hours and it then went into.

Q: Think about your first cup of tea, what kind of tea was it?

A: Normal Tea *click*

Q: What did you put in it?

A: Milk and sugar *click*

Q: Was it caffeinated of de-caffeinated?

A: Caffeinated *click*

Q: What did you drink it from?

A: A mug *click*

Q: How much of the mug did you drink?

A: All of it *click*

Q: Think about you second cup of tea, what kind of tea was it?

A: Normal Tea *click*

Q: What did you put in it?

A: Milk and sugar *click*
….

I think you can see where this is going, I had already answered two questions about the glasses of water I’d had (tap water, from a tall glass, room temperature, drank all of it) and I got as far as “Think about your third cup of tea” before I thought bugger this and left.

So think about what you’re asking, is it really needed? Can you make it quicker and easier for the user? For the example above change it to how do you normally drink your tea? and then did you drink any differently in the past 24 hours? Or a table where I could have answered each question at once for all cups of tea.

What did you put in your tea?
	Cup 1	Cup 2	Cup 3	Cup 4
Milk
Milk & Sugar
Sugar
Cream
Something else

2. Don’t ask unnecessary questions

Similar to the tea example above but for this is more about using the information you’ve already been given.

For example, from a mobile phone usage survey

Q: What is the make and model of your phone?

A: Apple IPhone

Q: Does your phone have Internet browsing capabilities?

A: Erm, yes, it’s an iPhone

Q: Can you play music on your phone?

A: Well yes, it’s like an iPod as well

Whilst I appreciate that some survey software won’t allow you to structure your questions this way just try to remove as much of it as possible if you can.

It’s well known best practice that people are less likely to fill in a sign up form if there are lots of fields, so use that same principle on your surveys and remove what you don’t need.

3. Make sure your options are sensible

How many times have you looked at a multi choice question and thought none of the answers fit.

Q: What do you think of first thing in the morning?

1 – Murdering children

2 – Murdering old people

3 – Murdering pets

You should really have a none of these/other type option all the time with a little box to write in an explanation. Yes it makes it harder to go through the results but it will make your results more accurate.
And none of this closest answer stuff either as a get out, you may think you’ve covered all angles and people couldn’t think of something else, but believe me, someone will.

4. Be clear in what you want

Ok this one came from another drinking survey, this time about alcohol. Now initially it referred to beer which for me is well, bitter and the like, lager is different. But they classed a few lagers in their example so I thought “oh they mean a more general term for beer”. Then a few questions on  “now thinking about Lager as opposed to Beer” and I thought well I’ll be damned if I’m off to change all my previous answers because you weren’t clear and I was thinking about Carlsberg and not John Smiths and so I left the survey unfinished.

As a basic rule if something can be misunderstood, it’s probably going to be so try to be as obvious and straight forward as possible.

5. Check the capabilities of the computer at the start

I got 90% through a survey once, I was using Firefox and the next question popped up and said please watch this video. Firefox had a moan about some windows media player plugin being needed and I couldn’t watch it, I couldn’t complete the survey and it wouldn’t let me restart it.

I should never have gotten 90% in, one of the first screens should have said “you will have to watch a video in this questionairre, is this one here working?”  and I would have sorted it out there and then before wasting half an hour on a survey I couldn’t complete.

6. Have a honest progress bar

I don’t know, I guess Microsoft are to blame for this one. For years we’ve had download progress bars that have seemingly done nothing other than print a random number on the screen.

“90 minutes to go”

“84 minutes to go”

“5 years to go”

“106 minutes to go”

So some people thought they didn’t have to make their progress bar accurate on their surveys.

Partly the problem comes from the fact that the progress bar is often based on how many pages of the survey you’ve completed. So an 8 page survey, you’ve done 4, you’ve completed 50%.

The only thing is pages 1-6 have 2 questions per page and pages 7 and 8 have 100 each! So the easiest way to fix it is to space out the questions as evenly as possible.

Hopefully these are errors which you’ll avoid in the future, feel free to add your own tips in the comments below!