AveryProg

googledgoogleeThere’s an article currently on the Rec List, which quite like about the murder suicide rate in the US. It’s an excellent diary.
However, it does one thing at the end of the diary, which is very common all over the web: It trusts the top line number of search results found by a Google search as accurate, and uses it as data to help prove a point.
That would be fine, except that number is complete fiction (i.e., Google search results number is completely made up.)
I don’t point this out to minimize the topic, or criticize the diarist in any way. But I do want to make sure, that as a reality based community, we’re using good data.
googlejgoogleigooglehI have done work for Google Search, and was told as much flat out. This was shocking, to say the least. I have always trusted that number, and it’s always so BIG! Somehow knowing that there were ten million results for “laughing goat” gave me perspective on my place in the universe.
Unfortunately, that top line number is made up out of thin air. Well, more likely it’s based on a complex algorithm probably using a statistical extrapolation of the number of results found. However, in any case, it’s always wrong. By a lot.
I once blew someone’s mind at a dinner party who just finished writing his sociology thesis based on the number of Google results found for specific Google Search result totals.
Oops.

1 thought on “AveryProg

  1. shinichi Post author

    Google Search Results Number Is Completely Made Up.

    by AveryProg

    Daily Kos

    http://www.dailykos.com/story/2013/01/11/1178286/-Google-Search-Results-Number-Is-Completely-Made-Up

    There’s an article currently on the Rec List, which quite like about the murder suicide rate in the US. It’s an excellent diary.

    However, it does one thing at the end of the diary, which is very common all over the web: It trusts the top line number of search results found by a Google search as accurate, and uses it as data to help prove a point.

    That would be fine, except that number is complete fiction.

    I don’t point this out to minimize the topic, or criticize the diarist in any way. But I do want to make sure, that as a reality based community, we’re using good data.

    I have done work for Google Search, and was told as much flat out. This was shocking, to say the least. I have always trusted that number, and it’s always so BIG! Somehow knowing that there were ten million results for “laughing goat” gave me perspective on my place in the universe.

    Unfortunately, that top line number is made up out of thin air. Well, more likely it’s based on a complex algorithm probably using a statistical extrapolation of the number of results found. However, in any case, it’s always wrong. By a lot.

    For example google this:

          masterpiece “phillips head” llama

    “44,000″ found. Go to page 16.

    Tricked you! There is no page 16. Only 143 were actually found. That’s an error rate of something on the order of 30,700%

    Please don’t use these numbers thinking that they mean anything specific. They don’t.

    I once blew someone’s mind at a dinner party who just finished writing his sociology thesis based on the number of Google results found for specific Google Search result totals.

    Oops.

    Reply

Leave a Reply

Your email address will not be published.