Google is a wonderful invention.  For most academics, google has transformed the way we conduct our research.  Finding information was never this fast.  With the emergence of scholar.google.com and books.google.com, it becomes much easier to find recent articles and track citations.

But as anyone who has spent five minutes in a classroom knows, google can misinform as well as inform, and Representative Bob Schaefer (R-Nampa) needs to refresher course.  According to published reports, Representative Schaefer, prior to debate on a voter ID bill, did a google search on the phrase “voter fraud” and came up with more than 900,000 hits.  Proof, the representative claimed, the voter fraud is rampant.

Well, no, it doesn’t quite work that way.  A search for “Santa Claus” comes up with more than sixteen million hits.  Virginia would be pleased.  ET no longer needs to phone home–searching on his name returns over three million hits.  Unicorns exist–15,700,000 hits!  Apparently we actually are in “The Matrix”, at least 40,900,000 websites think so.

Would it be impolitic to note that Google has nearly 900,000 hits on the phrase “Idaho skinheads”, coincidentally virtually identical to “voter fraud”?

Even a bit of attention to the links that Google returned should have caused the Representative some concern–six of the first twenty hits are stories or websites that argue that claims of voter fraud are massively overblown!

Counting “hits” as a quantitative indicator does have an intellectual lineage.  On newspaper databases such as Lexis-Nexis, scholars count the number of “hits” to measure the relative prominence of one news story versus another.  Content analysis of these same stories can help to study media bias, reporting for instance how frequently certain adjectives (incompetent) or verbs (resign)  occur nearby a proper noun (David Patterson).

But Google?  Google indexes a vast territory of information, some accurate, much inaccurate, and a lot outright junk.  At best, I suppose you could use hit rates over time on a phrase like “voter fraud” to index the prominence of this concern, and I suspect that the Representative is correct that more people have become more concerned about this over time.

But to use this as an indicator of actual occurrences of voter fraud is, well, intellectually fraudulent!