[Donald Rumsfeld Fan]

There is absolutely no statistically-valid conclusion you can draw from an internet poll, other than that the specific people who chose to vote in that specific poll probably have the preferences they stated.

You can draw the conclusion that people who get the majority of their news/analysis from the internet favor Fred ...can you not?

[Politicalities]

You can draw the conclusion that people who get the majority of their news/analysis from the internet favor Fred ...can you not?

No, you can not... unless you can prove that a) members of the population "people who get the majority of their news/analysis from the Internet" are the only respondents; a single respondent not in your population taints the sample, and b) that the subpopulation "people who get the majority of their news/analysis from the Internet and choose to fill out Internet polls, and happened to be visiting this particular site and voting in this particular poll" has exactly the same preferences as the full population. Both of these things must be true for you to draw that conclusion, and as it happens, both of them are false.

The More You Know!