quick bit of advice needed: incentives for completing questionnaires

[dodi1978]

Hi,

I don't often use surveys for research, but this time round I am. I need a quick bit of advice about incentives as I've got a bit of funding to pay some.

I am hoping for about 30 survey responses (more would be great, but as this is a pilot study without any advanced statistical analysis planned it doesn't really matter).

I could pay £5 (Amazon voucher) x 30 (but there'd be the danger of getting more than 30 responses - do I just make it clear in the information sheet that only the first 30 respondents get one)?

Or I could make it a lottery of 5 prices x £10 Amazon voucher (which would also save me some money that I could use on another part of the project - but how do I pick the winner?)

Anyway, I'd appreciate your advice on this!

The problem with advertising that you are offering £5 only to the first 30 respondents is that it can prove to actually be a disincentive if people think you will probably have already got 30 respondents by the time they come to reply. Some will assume you only want 30 respondents and that they are probably too late to be of value, while others will think that they won't respond for free while others before them have been paid. You run the risk of not even getting 30 respondents as people cannot accurately assess their likelihood of being in the first 30 because they will not know how many people you have invited or how quickly other people will have been to respond.

For the lottery / prize draw option, you need to consider whether entry into a draw to win £10 is actually going to incentivise anyone to take part. I would guess that it isn't, as the uncertain chance to win what is to most people a fairly insignificant amount is not much of an incentive. If your incentive isn't actually going to incentivise anyone, you might as well save you money. Psychologically, it can be better to offer fewer chances to win a larger amount of money, e.g. one prize of £50.

To select draw winners, you can assign each respondent who has opted in to the draw (ethically, it must be an opt-in process) a number and then use one or two rounds of a random number generator to select the winners. It is good practice to have someone else observe you as you run the generator.

Make sure you have some attention check, bot filters or capcha.

My favourite bot filter question currently is:

"DO NOT click any of the answers below. Please only press send to go to the next question."

has this work been approved by an ethics committee? they might have opinions on payment by lottery.

there are lots of platforms (e.g. Qualtrics / Gorilla / Prolific ) which will let you recruit paid participants in surveys and restrict your participants to particular groups by demographics etc. If you are going to do bigger studies, do look into using those.

we have had major trouble with Amazon vouchers - we paid ppt with amazon vouchers & then had multiple complaints of 'my voucher didn't work, can you send another' which might have been fraud but we couldn't prove it and had to pay out again. So be cautious.

Also, how long does your survey take? you need to pay at least £7.50 per hour and ideally £10 per hour.

Dear all, many thanks for your answers! I have now largely gone with the suggestions by the first respondent - 3 x £30 voucher by raffle. Hoping to avoid any of the Amazon problems mentioned here!
I had previous favourable FEO, but there were many changes to the survey and the way they are distributed, so I had to put in an amendment anyway - the rewards things was part of that. We'll see what Ethics say. It they are not happy with it, I am going to make changes.