random number /ticket generator free draw thing

[hargi]

Got a tricky one. Have a project that involves an anonymous questionnaire been filled in. We can not collect any person information at all from this.

But to encourage people to fill this in, if they do they will be entered into a free draw.

Now I dunno what solution to use for this.
If i could collect there email addresses then this could work, but we are not allowed to.

The other way is to make a random number at the end of the questionnaire, but then we have to keep a note of all the numbers generated and then we need some way of proving that that person got that number.

How would you go about solving this?

 

[visibleman]

As you say, the only logical solution would be to generate a reference number but it's an arse-way of doing things. What's the reason you can't store an individuals name/email address?

 

[UncleRuckus]

Seems completely backwards; why won't they allow you to email the winner?

 

[hargi]

Its supposed to be an anonymous questionnaire, they think people wont fill it in if they have there names attached to it or they may not answer it honestly.

 

[Staffy]

So how are they going to let the winner know?

Usually with these kind of things, you give them the option of entering the draw by filling in the email address.

 

[UncleRuckus]

Its supposed to be an anonymous questionnaire, they think people wont fill it in if they have there names attached to it or they may not answer it honestly.

An email isn't a persons name though. I'd fire the client.

 

[UncleRuckus]

Also I wanted to add something a bit more constructive.

If a person is not adding an email they could fill out the questionnaire multiple times with false data just to give themselves a higher chance of winning the prize.

Which seem to me kinda of counter productive for market research, which I assume you are doing?

 

[durbs]

You need an email to do this. Why not make it optional, say at the end "if you want to entered into the £100 prize draw, enter your mail here" along with assurances the data will not be misused.

That way you can still advertise the survey as "£100 prize!" but the onus is on the user to take up that offer once you have the data. If they don't want to and wish to remain anonymous then fair enough, you've got what you need from them already.

No other clean way to do it really.

 

[AHarvey]

could you not create a cookie that generates a unique ID for them and simply tell them at the end of the questionnaire to come back between dates X and Y and if their cookie matches the winning number they win a prize?

knowledge of cookies is limited but one per PC so doesn't matter how many times they fill it in?

 

[UncleRuckus]

could you not create a cookie that generates a unique ID for them and simply tell them at the end of the questionnaire to come back between dates X and Y and if their cookie matches the winning number they win a prize?

knowledge of cookies is limited but one per PC so doesn't matter how many times they fill it in?

What you've got to take into account here is that: people clear their cookies or have them disabled all together.

Plus it wouldn't be hard to hack a script together which increments a number stored in the cookie and refreshes the page until you get the winning number.

 

[hargi]

This is true, the survey is a college survey to find out about the students behaviour and lifestyle. Do they smoke, drink take drugs etc? This is to allow the relevant people to come in and deliver training and get funding etc. The data collected should not be able to be used to identify a individual student.
Students can do the survey from anywhere, but the client thinks that most will be doing it in the college, so individual pc cookies will not work.

The easiest way is to get them to opt in and enter an email address.

The only other way i can think of is to generate 2 random numbers per finished survey get the student to note them both. Then just publish the first number on the website/noticeboard and get the student to provide the other number to provide proof. But this seems very backward

 

[Staffy]

Email collection as a option is the best way to be honest. The random number idea is good but you would have to make sure there are plenty of safeguards put in place.

 

[sunama]

The only other way i can think of is to generate 2 random numbers per finished survey get the student to note them both. Then just publish the first number on the website/noticeboard and get the student to provide the other number to provide proof. But this seems very backward

This is a brilliant idea.
Make sure that the random numbers you generate are very long.

Also, lets say there is no email address involved, how are you going to award the prize to the person?

For example.
1. I fill in a questionnaire.
2. At the end of it, I get my prize draw numbers.
3. At that same moment both numbers are stored on your server.
4. The prize draw is made and both winning numbers associated with the winner and noted. The first number is posted on a notice board.

What happens now? How do I claim my prize? Surely, I now have to come out of the shadows and reveal my identity...which defeats the original solution which gave the entrant, anonymity. Correct?

Unless you can get around this problem, it would probably be best to allow the entrant to choose to enter the draw (and provide email addy) OR not provide email addy and not enter the draw.

I can't see how anybody can remain totally anonymous, if they are going to enter the draw. At some stage, the winner will have to reveal themselves.

If you want, you can really turn this into a game and leave a cheque in a safe, whose combination corresponds to 2nd number which the entrant has. This would guarantee complete anonymity. But IMO...this will become a little silly.

 

[visibleman]

As said numerous times, it's completely backwards having a competition without any method of contacting the winner. Simple solution - Make the draw opt-in and separate the draw from the survey ie: store the draw email addresses in another table, obviously highlight this fact to the users/students, that way there isn't a direct connection between results and emails addresses.

 

[AHarvey]

Another idea, get it to give them a 12 digit code, keep a reference for yourself to say if it has been used, then get them to send their entries in, or drop them in a collection box in the reception with their details on the back.

Randomly pick a number that has been used and there you go.

You are still collecting data, not collecting personal info to match the questionnaire and can give a prize away

 

[lookitsjonno]

#1 how about a phone number instead? much harder to identify someone from a phone number.

#2 Get them to print or save a completion certificate which contains an MD5 hash of their answers and a 3 digit code. You then publish winning code and once winner contacts you you compare the MD5 hash with their answers to distinguish if the certificate is genuine.

 

[sniffy]

I'd argue for an opt-in for the prize draw via email and make it very clear that the survey data and email addresses are in no way linked. This is far more convenient for the user and assuming you will only allow addresses from a respective domain you can take away any incentive gained from multiple submissions (which will otherwise be a problem as you're dealing with students). If that simply isn't an option then two randomly generated numbers given to the user after survey completion would work. If you wish to offer the incentive but decrease the chance it gets collected then I imagine this is actually preferred.

 

[Mark M]

How about a link at the end of the survey to a septet prize draw? You can only enter the prize draw by finishing the survey and stress it's not related to the anonymous survey and no data can be linked.

 

[dave_a141]

How about a link at the end of the survey to a septet prize draw? You can only enter the prize draw by finishing the survey and stress it's not related to the anonymous survey and no data can be linked.

Exactly what I was about to say. In your previous post you said that the data mustn't be able to identify the user - if the email address and survey data is stored separately then there is no way to do so