I agree with all that has been said so far on how you could end up $1000 gone with no real data.

As you have not specified a criteria for these people but you do not need to know them or collect any data about them and they would be selected randomly by tear off tabs then I would suggest that you engage a panel company to recruit the data subjects for you.

This would mean you would not get any personal data, the panel company would handle the payouts and you could specify a minimum target for the people you want (age, gender, employment status etc.)

The way this would work is that you pay the panel company the cash they direct the respondents to your survey and you redirect the respondents at the end of the survey back to the panel company so they register as a complete .... you can also do this with screenouts as well (using the quota system).

All panel companies will give you a CPI (cost per interview) dependent on what your screening criteria and interview length is ... and the majority of respondents will cost you less than $10.
.
This way you have a survey with anonymous people and a way to protect your $1000

Hi Dave, I was also thinking about a panel before. However, reading about the very specific target group, I highly doubt that any online panel will be able to help here (specific sport, specific region, etc.).

holch wrote: OK, so this is complete self selection and you would need to provide many, many tokens (and thus amazon codes) to make sure you reach your n=100 respondents. So this is definitely not an option. If you would only give out n=100 tokens, this could work, but then you would have to have a very defined sample.

So...technically a bad bad bad idea...

I was reading about tokens in the manual and in the forums and it seems like if I gave the tokens on the tear tabs I would create a dummy participant table...makes sense...

But I wonder if there's a way to have them "sign up" instead so I can send the code to them through the email function on LimeSurvey...without me seeing their email address or name...

Either way, I'm going to meet with my supervisor and bring up all of these concerns.

You can't make them signup and send them the codes without seeing the email. This is not implemented in Limesurvey.

I honestly see no problem if you see their email address. This is done in market research all the time. If you disconnect their personal data from their responses, you have no way to connect them to their results and that is what counts.

Complete technical anonymity is nice, but comes with a lot of disadvantages and makes some things impossible or very dangerous (like in your case).

What you could do is to let them participate in the survey and then via the end url feature you redirect them to another survey. If you don't pass on any code, there is no way for you to connect their responses to the second survey, where you ask them for their contact details to send them the code. It doesn't protect you completely from people answering more than once to get an additional voucher, but at least it makes it harder. If you then also use tokens with your link, they would have to take more than one paper with them to be able to respond more than once, or go back and get a others. Overall, the recruitment approach together with incentives is always going to be somehow tricky.

For a scientific project I honestly don't see any need to provide incentives at all. People usually participate because they are interested in the study to be a success or because they want to help out the scientist. If your survey isn't extremely long, of course.

holch wrote:
What you could do is to let them participate in the survey and then via the end url feature you redirect them to another survey. If you don't pass on any code, there is no way for you to connect their responses to the second survey, where you ask them for their contact details to send them the code.

Thanks, holch

That seems like the most viable solution.

Redirect to a survey with 1 question that says "please provide an email address to which we can send your $10" and then from that, I can compile the email addresses and just send an email to each like:

"Dear participant,

blah blah blah, here's a code for $10 to amazon.

111 222 333 444

Thanks for your participation,

Researcher"

And one tip from market research practice: I have worked with raffles of 50 Euro vouchers (20 per 1000 participants) for years for many, many surveys. We purchased many vouchers each year and you'll be surprised how many people will NOT redeem their vouchers. At least I was.

I wouldn't want to give presents to Amazon, they have enough money. So what we did was taking the expiration date of the vouchers purchased and take off one month and communicate this expiration date to the respondents. I think the expiration date is somewhere around 1 year.

So you give them 11 months. Then when the date is over that you gave to the respondents, you check if the vouchers have been redeemed. If not, you redeem them. We increased the library in the company quite a bit and could also buy quite a few appliances with it as well.
No one will come after eleven months and say, ohhh, wait, my voucher! And if they do you can just pay them their 10 dollars. In various years we had no case and actually the respondents can't even complain, because the voucher for them has actually expired. I just wouldn't want give money to Amazon for nothing. Jeff doesn't need more money. He is fine, even after the divorce.

holch wrote: What you could do is to let them participate in the survey and then via the end url feature you redirect them to another survey. If you don't pass on any code, there is no way for you to connect their responses to the second survey

Without passing any additional code, you risk that the URL of the second survey is directly accessed and distributed via social media. That is the worst case, cause you end up with a lot of responses and don't know which is valid. Even if you have all email addresses of potential valid respondents, you cannot expect that people use the same emailadresss for the vouchers.

LimeSurvey is missing a questiontype for collecting contact data separately.
E.g. like this www.soscisurvey.de/help/doku.php/en:create:questions:contact

As I said, the combination is not perfect and someone savvy will find a way to get to the vouchers. The solution is a lot saver, but yes, it would be best to have a question within the survey and then you just separate the two data sets. Unfortunately LS does not have a feature to keep the personal data out of the results table, but then anonymity is generally not only a technical thing, but rather a process. For years market research has run anonymous surveys, that depended purely on how the data were treated, because if you wanted to, you COULD connect personal data with responses. A responsible company wouldn't do it. But you could.