Take the 2-minute tour ×
Super User is a question and answer site for computer enthusiasts and power users. It's 100% free, no registration required.

I voted on a website ( not mine), on certain issue yesterday. The result was one sided ( i.e., more than 90% of the people answered Yes at the time of voting, about 9k people had voted on a one week old poll at that time). However today when I checked the result, only 50% of the people answered Yes. How can it that in one day more than 9k people voted when the average number of people who voted in past few days was only 1k?

Is there anyway to check whether that third party online poll has been tempered with? I don't have access to logs, and I don't closely monitor the "time series" of the votes. But still, is there any other ways to check this?

share|improve this question
7  
I'd just assume that all online polls have been tampered with. Safer that way. –  Grant Aug 11 '09 at 11:24

4 Answers 4

up vote 12 down vote accepted

Do not assume that any online poll is valid.

share|improve this answer
1  
Trust online polls as far as you can throw them. Don't forget that the servers are bolted to the floors in a remote location. –  Grant Aug 11 '09 at 11:27
    
Well, there are polls where you need to be a registered user (and pay money). Those I'd trust as far as I could throw the server if it were unbolted and I could reach the server. –  Daniel H Aug 11 '09 at 14:32

This is non-trivial even for the owner of the site that run the poll.

For you as an outsider, forget it. There is no way you can verify if the votes are real. As with most information on the Internet, it comes down to trust.

share|improve this answer

Tampering is hard enough to spot if you have full access to the server, from the outside it would be almost impossible.

But there's a wider issue than tampering.

Even if you could 100% guarantee that the poll hadn't been tampered with it still wouldn't be valid.

Think of it this way - put the same poll about Linux vs. Windows (say "Which is better, Linux or Windows") on Slashdot and on Microsoft.com. Do you think you'd get the same results? Of course not.

The reason are (and this pretty much applies to all on-line polls other than those done by polling organisations on an invite basis):

1) The group you're polling is self selected. Not everyone will have chosen to vote therefore the opinion of those in the middle ground are either missed or under representated as those with extreme opinions are more likely to vote

2) It will have an intrinsic bias based on the site doing the poll (as in the above example). A site might appeal to those with a particular view or just to a particular demographic who are not representative of the world as a whole. Which leads to...

3) The results won't be weighted to make it representative of the whole group in question. So if your website is popular among men but you're interested in opinion of the population as a whole you need to balance that so the votes of men represent only 50% of the whole. How is that achieved in an on-line poll? Simply it isn't.

4) (Thanks to @TFM) The question can be biased. For example questions with a negative encouragement: "Would you consider Linux even if it's complicated to install?"

Essentially on-line polls will tell you nothing you can rely on. If they are right it's as likely to be luck as anything else.

share|improve this answer
    
4) The question might be biased. For example questions with a negative encouragement: "Would you consider Linux even if it's complicated to install?". –  TFM Aug 11 '09 at 12:04

What does "untempered" even mean? Through twitter and co it's possible to drive a massive amount of people to page who will vote in a predecided way and therefore destroy any representativity of the poll.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.