Rock musician Trent Reznor wasn't the only famous person who recently conducted a survey. Scott Adams, creator of the well-known comic strip "Dilbert," decided he wanted some "useful and unbiased information about which candidate has the best plans for the economy" and decided the best way to get that information was to ask a bunch of economists.
So Adams commissioned a survey of over 500 economists who are members of the American Economic Association. You can see a nice slideshow presentation of the survey results here (if you want the short version of the results, click over to slide 89 to get the Summary of Key Findings), and you can get a complete copy of the survey report by sending an email to email@example.com. [Interestingly, while the company that actually conducted the survey, The OSR Group, is apparently a "respected national public opinion and marketing research company," they don't seem to have a website.]
You can decide for yourself how useful the findings of this survey actually are (the group generally felt Obama would be better than McCain for the economy by about a 60%-30% margin, with the remaining 10% seeing no difference between the candidates). I personally was most surprised to discover that nationally, most economists are not only male but most are also registered Democrats; the survey sample attempted to reflect this imbalance, with 48% of the sample being Democrats, 17% Republicans, and 27% Independents.
But the most interesting thing to me is the fact that someone not in academia or marketing or a social science research-related field would conduct a survey to gather information about some topic. While all of us are invited to take surveys of one kind or another on an almost daily basis, how many of us would ever think of conducting one ourselves, or hiring someone to conduct one on our behalf?
While surveys like the ones Adams and Reznor conducted are fairly involved and require a good bit of time and money, both to conduct the survey and to analyze the results (Adams notes that he undertook his survey "at considerable personal expense"), there are many types of surveys that can be conducted with much less time and effort, especially if it's not critical that your results be generalizable to a larger population or achieve a particular level of statistical significance. Many small companies do it all the time: they ask their customers what they think of potential product offerings, or they ask the people who just participated in a seminar what they found useful about it. Often a marketing and research company (or an evaluation consultant) can help you determine if a survey would be useful for you, and even help with its design, implementation, and analysis of the results.