Google Trends: an accurate way to gauge interest in political candidates?
People use Google Search every day to look for whatever they want to learn more about. Google tracks the aggregate searches of its countless users and compiles them in their Google Trends database based on search volume data. Here, topics are given a score from 0 to 100 based on how often users search it over time and compared to other topics. This can be very useful during periods of increased interest, such as the Democratic Debate. Google Trends attempts to track who “wins” debates based on how many searches occurred during the 3-hour time period for each candidate. Google Trends doesn’t always line up with polling however; Marianne Williamson, for example, had some of the most searches (because of her bizarre candidacy and positions) but still had low polling ratings. Another issue for Google Trends is that it doesn’t show which sites people are going to click on.
When search engines receive a search query, they must present the pages in a certain order. Generally, search engines want to have the most relevant pages at the top of the list, so that people click on a link, and they receive ad revenue. Currently, Google Trends communicates what is being searched, but not the page rank of what is shown in the search. If they combine the data with the page ranks of the searches, perhaps they could alter the page ranks for specific time periods of peak interest, or utilize similar Trends as a part of the algorithm. For example, if two candidates spike at the same time in searches, related pages could pop up in the page rank that may not be otherwise connected through links.