Estimating dominance centered on Yahoo queries: Why it is an awful idea
Many people lookup the internet for a couple of subject areas and you can upcoming use the level of serp's ("hits") for each and every point to rank the fresh new relative popularity of the information. On 2011 Mutual Mathematical Meetings (JSM), I'd the ability to sit in numerous conversations of the statisticians out-of Bing or any other large Internet sites people. Once i chatted with some of those statisticians shortly after talks, they verified the thing i had thought: it's an awful idea in order to imagine the rise in popularity of men or device according to the result of an on-line lookup.
An instance data: Sizzling hot pet instead of hamburgers
Easily choose "very hot dogs," the search engines informs me discover "about twenty-six,700,000 show." If i identify "burgers," I find that there exists "in the 20,900,000 efficiency." Not simply exactly how many performance, but in addition the amount of Sites lookups favor "scorching pets" more "hamburgers". Would it be good to conclude one hot pets are more popular than burgers? You can find out of the exploring statistics which can be linked to practices.