Mark Bauerlein counted citations by using Google Scholar, which he argues is the "best instrument" (9) for tracking citations in humanities fields (for which there are currently no reliable references at all). This strikes me as not, perhaps, the best method. To explain why, I'll use myself as an example.
Problem #1: The existence of one's articles in the MLA database.
Four of my articles are nowhere to be found. (You can see a full list of my main publications on this blog's "about" page.) One of them is intended for use by acquisitions librarians (the bibliography I did for Choice), not for scholarship per se, although I've come across at least one person citing it as such. Three others that are nowhere to be found in the database: Grace Aguilar, Emily Sarah Holt, and Anne Boleyn. The first two are in book collections; the last is in a journal that the MLA indexes in theory, but apparently not in practice.
For obvious reasons, it's hard to cite something if you don't know it's there, although a couple of people have stumbled over Grace and Anne. All of the articles missing from the MLA are present in other online indexes, however.
Problem #2: Using Google Scholar to find citations.
Although Google Scholar pulled two citations of my work from Project Muse, it missed the one in JSTOR. It also missed a citation in PQ and, more importantly, something on the order of eleven or twelve book citations. This is a rather large margin of error. (There are also some dissertations that don't get pulled up on Google Scholar.) Bear in mind that I'm simply counting, not evaluating the quality of the citation, or how my work was used.
To track citations of my work, I have to do the following:
1) Check Project Muse, JSTOR, Academic Search Complete, Humanities International Complete, and the Taylor & Francis Archive--directly. All of these searches pull up overlapping but different results (not all of them appearing in Google Scholar).
2) Check Google Scholar.
3) Check GoogleBooks.
4) Check Amazon's and Amazon.co.uk's "search inside the book feature" (this doesn't work very well).
5) Check Google.
Moreover, if I'm tracking myself by name, I have to run the following searches:
Miriam Elizabeth Burstein (I normally publish using my full name)
Miriam Burstein (because people sometimes just ignore the middle name)
M. E. Burstein (because house style occasionally mandates initials)
Miriam Burnstein or Miriam Elizabeth Burnstein (yes, there are multiple citations out there with my last name misspelled; when I tell people that I use my middle name because "it's the only one people can spell correctly," I'm not entirely joking)
This is all for one person!
Moreover, all of this presumes digitization. Many humanities journals just aren't available through either JSTOR or Project Muse, or can't be searched without a subscription. Books are similarly subject to the vagaries of their publishers, as they may be unsearchable in GoogleBooks, and Amazon's search engine does an apparently random job of doing keyword searches in books with "search inside the book" enabled (it will pull up a handful of books in which I'm cited, but miss many others, even if the book itself can be searched once you've loaded it into the viewer).