Question on surveys

[nealrm]

I have a survey on my site so the users can provide feedback on the new layout. I have noticed an odd trend in the response rate for the non-US users. While non-US users account for about 0.5% of all users, they account for about 10% of the survey responses. Any ideas on why this might be?

[vangogh]

Could always be coincidence. I'm not sure if there are any cultural differences where Americans are less likely to want to be helpful than our non American counterparts, but I suppose that could also explain things.

Do you suspect some of the respondents are automated? Maybe thinking they get a link with the vote or some other benefit?

How many responses are you talking total? Is it possible that you haven't gotten enough responses for this to be statistically significant. I've run a few polls where the results at the end were vastly different from the results at the beginning.

Is the poll visible on all pages? If it's only visible on a page or a few pages maybe it's possible that page ranks better in non U.S. search engines or maybe the page(s) with the poll happen to be pulling in more non U.S. traffic. The % you mention might be for the site as a whole, but maybe the % for the page(s) in question might be different.