Google’s new “reading level” filtering
Google has added a feature to its advanced search form that allows you to filter results by reading level or add information about a page’s reading level to the information in the results. Reading level is indicated as “basic,” “intermediate,” or “advanced.” Like most of what goes on underneath the Google hood, we aren’t given much information about how reading level is computed.
I am constitutionally against anything that could be construed as “dumbing down,” but I have to confess that I find this feature interesting. Working with first-year students in an academic library I often find myself wishing that we had a way to search bibliographic databases that would provide scholarly acceptable content that the students were actually able to comprehend. Something like this technology could be used in a bibliographic database, although I am sure its application in a reference setting would be potentially awkward and intellectual freedom issues would emerge.
In checking out this feature, I noticed that Google’s advanced search page includes some additions that I would have to call welcome and surprising from a librarian’s standpoint. If you haven’t looked at it for a while you should check it out (including the collapsed features at the bottom).
2 comments on “Google’s new “reading level” filtering”
Off the cuff, I think this could be really helpful for lower level readers, such as children. When a child Googles “Zeus” (or even “Greek god Zeus”) for their assignment, not only are there likely to be a lot of irrelevant results, but are also likely to be a lot of results that are above their reading level. This could help them sort out sources that will be usable.
Of course, a librarian has other ways of finding appropriate sources, but a child doesn’t always have access to one.
Alas, automated (read, mathematical) algorithms used to measure reading levels are nothing new. See the Flesh Readability Score (http://bit.ly/emZq8R) as an example.
With the advent of so much full text is more than possible to calculate such scores and use them as a part of an item’s descriptive metadata, and thus incorporate them into search interfaces.
I have done this a bit on my own. For example, you can look at my description of Henry David Thoreau’s Walden (http://bit.ly/hwNzGK) and see that it has a Readability Score of 55, average.
Comments are closed.