Tuesday, October 13, 2009

Grading with Voice Recognition Software

Like most people, I've always done my grading by hand--students hand in a physical paper, I take a stack of them to a cafe or the library, and settle down with my pen. Again like most people, I try to provide useful comments that will help the students in the future.

This quarter I've been trying out the electronic dropbox, where students have to turn the work in online. My main reason for this was admittedly that I wanted to try out the anti-plagiarism software Turnitin, and that requires digital papers. (So far there has been no hint of plagiarism, I am glad to say.) My secondary reason was that I thought it would be worth trying typed comments. My handwriting is not terrible, but my hand does get tired and sometimes I do spill coffee on papers. I've been using Word's tracking and commenting functions. It seems to work reasonably well, except that I find I experience much more eyestrain.

James M. Lang's recent article in the Chronicle of Higher Education provides an interesting look at a rather different approach to grading papers electronically--the use of voice recognition software. Lang's colleague, Geoffrey Vaughan, uses Dragon Naturally Speaking to compose responses to papers. Lang states:
"[F]or years now I have been typing up my remarks on students' work. I began doing that because my handwriting is so illegible, but quickly found it had two unforeseen benefits. First, because I can type faster than I can write, I can give more substantial comments on each paper without adding additional grading time. But second, and more important, instead of my students flipping right to the back page of their paper to see the grade and comments together, my students now almost always sit and read my typed comments first (they are stapled to the front of the paper) and only then flip to the end and check the grade. It may be a small change in how they process the graded work, but to me it sends the right message that the written feedback matters more than the grade."

Intrigued by Vaughan's twist on typed comments, Lang asked for a demonstration of how he dealt with student work using Dragon. Vaughan showed how he gave "a full page of text, single spaced, in numbered paragraphs. Each paragraph corresponded to a number he had placed in the margins of the student paper. A short paragraph at the end provides the standard final comment."

Vaughan says
"I used to put some scribbling in the margins of each paper, and I would draw lines from the margin to the bottom or write 'See over' and then write two or three sentences of a final comment at the end of the paper. Now I simply put a number in the margin, and I speak my response to each of those numbered points."

Lang observes, "His paragraphs are conversational, as you might expect, and not perfectly formulated. Vaughn explains to his students that he uses the software, and that they can always come to him to clarify confusing elements in his comments." He notes that
"As we talked, and I read through the samples he had, it became apparent to me that Vaughan's real commitment to this mode of responding to papers relates more to the conversational tone of his responses than to the time-saving element of it. As he explained it to me, the software has inspired him to think about responding to students' work as more of a dialogue than a summary judgment—a model he learned during the year he spent as a graduate student and tutor at the University of Oxford."

During Vaughan's days at Oxford, he'd been accustomed to having students come and read their papers aloud; the tutor would periodically stop them and discuss their ideas as they read. Use of voice recognition software reminded him of that process and makes him feel like he's having a conversation with the student. "It combines the usual American-style, paper-grading process with the model of the Oxford tutorial," he says.

Vaughan's comments reminded Lang of what he finds "the main challenge of evaluating student work"--the two separate functions of the grading process--"to explain to the students the reasons for the grade they received (i.e., you did these things well, and these things poorly) and to help them understand how to improve their performance." Lang observes,
"To turn our response to a student's work into a dialogue in which we are not simply passing judgment but engaging in a conversation about how the student can improve seems like a pedagogical change worth adopting, whether or not it saves time."

Voice recognition software is pretty easy to use these days. I have Dragon's next-to-latest version and use it to dictate long quotations when taking notes. The key to using voice recognition is simply to spend enough time at the beginning getting it used to your voice and the type of vocabulary you use. It's important to make corrections to its mistakes using the software, because that helps train it not to make the same mistakes again.

I don't know whether I'll take to doing my paper comments this way--I am not good at dictating my thoughts--but this might be a great method for others to try.

1 comment:

Peter said...

I've been grading electronic assignments since last century and quickly worked out that there are ways to reduce the workload and save time while at the same time providing more detailed and useful feedback to students. One of the major ways it to use reusable comments that can be easily modified for each student. Probably 60% of comments on a paper will be reused and even this gives more time for personalised comments when they are needed. You can see video demonstration at:

The software also allows you to record a voice comments which is embedded in the word file so it can be easily returned to the student. This is shown about 4:20 into the video

A 30 day trial of of the software is available.

Peter Evans
eMarking Assistant