Mathematical Illiteracy in the Media

From a SCIENCE magazine, http://seedmagazine.com/content/print/a_writing_revolution/

Since 1400, book authorship has grown nearly tenfold in each century. Currently, authorship, including books and new media, is growing nearly tenfold each year. That’s 100 times faster.

So how does this happen?  How does a statement that's absurd on the face of it get published?  Did no one actually stop and think what it was saying?  This was published on October 20, 2009 and there has been no correction so far.

Thinking it might be a typo, I looked into their methodology.  They took estimates for the number of book and blog authors in various years (and Facebook and Twitter "authors") and fitted them to an exponential curve.  They then took the mean of the fitted parameters for the blog, Facebook and Twitter curves to form a new curve and since the slope of this new curve on a logarithmic scale is about 100 times that of the slope for the book authors curve, they conclude that new media is growing 100 times faster than books.

Now, I don't think the authors are stupid.  One is a psychology professor and the other is the inventor of Wingdings, so they must be smart enough guys.  I just don't understand why this wasn't caught before being published.  According to their own model, the growth at the start of 2009 was 9.8K book authors/year, 57M blog authors/year, 12M Facebook authors/year and 2.3M Twitter authors/year.

Anyway, if you have any other examples or advice, post away.

Dear Anonymous

I am an author of the Seed article. There seems to be a misunderstanding. As you note, we do say, "Since 1400, book authorship has grown nearly tenfold in each century. Currently, authorship, including books and new media, is growing nearly tenfold each year. That’s 100 times faster."

You say this is "absurd on the face of it". However, I wonder what you think we are saying. Tenfold growth means that authorship grew by a factor of ten. For books, tenfold growth takes a century. For new media, tenfold growth takes a year. The data on the graph show this. Do you disbelieve the data? Since a century is 100 years, it follows that tenfold growth occurs 100 times faster (takes 1/100 of the time) in new media as in books.

Perhaps you took "growth" to mean rate of increase (books/year). In our article, we made an effort to say things in a way that was correct, yet still accessible to readers who are unfamiliar with logarithms and exponential growth. As you know, exponential growth is best visualized with a logarithmic scale, because it is then displayed as a straight line. Exponential growth is characterized by the slope of that line, e.g. as factors of ten per year. That slope is 100 times higher for new media than for books. Does that help?
http://psych.nyu.edu/pelli/pubs/pelli2009writing-graph-data.pdf
http://psych.nyu.edu/pelli/pubs/pelli2009writing-graph-regression.pdf

We discuss the thorny question of how to define "publishing" and "authorship" in our reply to comments at the New York Times blog.
http://ideas.blogs.nytimes.com/2009/10/22/a-writing-revolution/#comment-28215