[jpeterman]

The only really valid part of the linked article is the point about how you can't make sweeping generalizations based on a very small sample size and a (relatively) small trend.

Concluding that people probably know what the EU is while searching very specifically using the phrase "What is the EU" is no more accurate or truthful than concluding that they don't know what the EU is.

The author says at one point, "perhaps we can more likely conclude". Perhaps we can. But perhaps we shouldn't make any conclusions and just stick to the thesis that the press should know better when reporting facts?

[mcbits]

True, the article gives no basis to conclude that people searching "what is the EU" already know what it is, but it seems like a fair default position. I might search "history of the EU" or "Mitt Romney bio" to get a certain kind of content outside of current events and gossip. I could also see other people with similar (limited, but much greater than 0) knowledge searching "what is the EU" or "who is Mitt Romney" for exactly the same reasons.

The less clueless-sounding queries have spikes on Google Trends that appear very correlated with the "who is"/"what is" spikes, for what it's worth.

[SixSigma]

Agreed. That's how to do sensible searching.

I didn't know how exit would happen via "Article 50", so I couldn't search for that term.

This election cycle has demonstrated that Google is acting politically, not a neutral repository with great search tools.