An interesting recent phenomenon is the popularity of internet polls. Traditionally polls and surveys have been conducted by agencies specializing in formal statistical analysis with emphasis on content. However, today virtually everyone seems to be conducting surveys, on a variety of topics ranging from opinions on what's wrong with the world to favorite James Bond actor. For these, statistical analysis does not appear to be the main goal (does anyone really care about the confidence interval on who was voted as the most popular James Bond actor?). The main reason for this may be the fact that on the internet it is difficult to derive a random, unbiased sample for any meaningful analysis. Consequently, the emphasis appears to be entertainment. But is there any value to internet polls beyond entertainment?

One interesting use for internet polls may be in gaining insight into the preferences of the audience by observing the participation rate (# of voters to # of visitors) over time on varying topics. The point of interest here is not the distribution of votes across the choices on a particular poll, but the possible trend over time between poll topics. For example, an observation made by a medical reference website may reveal that the response rates on polls related to infant care are consistently higher than geriatrics. This information may be helpful for deciding future content of articles and advertisements or for identifying areas of potential marketing research. Although this approach may not be the most accurate means for determining preferences, it appears to be a very cost effective tool for analysis beyond entertainment.

More Stats in Industry