http://subtleshift.blogspot.com/2008/06/sin-damnation-google-trends-and.html
Interesting post on Google Trend results for controversial religious Google search terms coming from South Africa. Seems like a combination of genuine controversy and statistical skewing, but very worth thinking about.
Thanks, enjoyed that post. I mostly agree with that blogger. Google Trends is only useful in certain situations. Still fun though ;D
The next post there lists even more search terms where we came up number 1 or 2. It’s actually quite funny, but I really want to know if it is due to genuine relative search frequencies, or just a statistical fluke. Any ideas? www.subtleshift.blogspot.com
I think it’s like he explained in the first blog post. Since people spend less time online in SA, they tend to search more for terms that they actually want to research - and not on frivolous everyday topics. So total searches tend to be less, and it is easier for a specific term to get a high normalized score in Google Trends.
In other countries that buy online a lot, the search term “psalms” will have to compete with terms like “xbox” or “hats” for example, so those countries won’t feature as high for “psalms” as SA might. Not to mention that these are just the english speaking countries, which cuts out quite a bit
What is also interesting, is that Philippines also feature quite prominently with South Africa…
But how does one then explain the results for “Sunday” and all the other days of the week? If South Africa and the Philippines are getting pushed up due to low search volumes, why are we not in the top searchers for all the days of the week? I was puzzled by the results for Tuesday, Wednesday and Friday, until I remembered Shrove Tuesday, Ash Wednesday and Good Friday, but SA being number 2 for Sunday certainly implies that we DO search for religious terms more often, doesn’t it?
There are also “sunday school”, “sunday lunch”, “sunday drive”, “sunday times” which will all skew the results and aren’t all religious - as I said before, Google Trends is interesting, but mostly useless as an indicator on any specific level.
Yes, that makes sense. I agree that the normalized country and city rankings are probably useless, but the shape of the volume graph certainly gives more useful indications. Just try Trending ‘xenophobia’, or ‘oil price’ and ‘electric car’.
Or even better, look at “Richard Dawkins” and compare with the release date for The God Delusion. And let’s not forget 666: A HUGE spike on 6 June 2006, when rumours spread like wildfire in Catholic countries about unbaptized children and whatnot.
Must admit, Trends has been my biggest time wasting activity in the last couple of days, but the novelty is wearing off.