Will We Ever Allow Computers to Grade Students' Writing?
Writing is taught as both an art and a science. As we go higher in the grades, student writing becomes longer and genres broaden. Teacher loads get higher as well. A high school English teacher could have 125 students or more. If unable to stagger class assignments, a teacher can have 125 (or more) papers to grade at one time. Practically, that can limit the number of writing assignments given. The less writing our students do, the more likely chances of their improvement are diminished. Enter technology.
"NO!" is the cry heard around the educational universe. "A computer cannot grade a paper as well as I!" Teachers make their opinions heard. Leaders listen. Case closed. We are not suggesting that a computer can replace the grace and value of the interaction between a teacher and a student in the development of writing as a communication skill. But let's have a discussion before we reach "Case closed."
First, truthfully, what is the chance that a teacher with 125 students can successfully meet with and coach their students through each piece of their writing? Secondly, who actually believes that the comments a teacher may write on a paper does more than direct corrections? Nothing replaces the teacher having a conversation with and coaching a student. Who wouldn't agree?
Sci-Fi and perhaps the media, and certainly our lack of real knowledge about data and technology have us believing computers can't grade student writing as well as a teacher can. But what if we opened our minds and began to think about the use of Big Data Analytics and its potential role in creating the space for teachers to maximize their talents in working directly with students to improve their performance? Big Data is a concept far beyond what anyone we know can honestly wrap their heads around but...we can understand what potential it holds. Imagine a computer software program, developed to analyze student writing, that has been written in response to the same assignment and graded by the same teacher. Imagine that all of this data, growing over time, year after year, could be analyzed and coded so that the program would allow for computerized evaluation of at least some aspect or aspects of the writing even if only during the drafting process. Although an admittedly small 'N' or sample, it is the work of that teacher and not normed by anyone but that teacher (or group of teachers). That would mean the basis for the analysis of the papers would be the teacher's own grading history, or the history of the group. On a larger scale what if, in larger schools, this program included all the teachers who required that assignment? A beginning to be safe and sure should be a local solution. But the possibilities for larger samples exist, once we came to trust the data.
There is some acknowledgement needed first. Teachers' grading of essays may be neither valid nor reliable, yet the measure is used and trusted. It may not be truly possible for a teacher to grade 125 papers, semester after semester and be consistent in the analysis of the students' work. Can the teacher bring the same set of standards to the analysis of student work over 125 papers, over time? Of course not.
Here is where leadership is needed from teachers and administrators. Before some company comes up with a way to grade essays and boards of education become enamored with the idea, and legislators find new ways to require their use...let's lead. The technology is here. Big Data Analytics are already being used to do things like sentiment analysis, which can take written work and analyze opinions, views, beliefs, and convictions (Zadrozny & Kodali. p.255). So why can't we come up with a way to exploit this technology in order to improve our current learning opportunities. Doesn't it makes sense that if the assessment of student work is based upon the teacher's own grading history, that it has the potential to become a trusted ally in the teaching and learning process? Too expensive to find a data specialist who knows how set up the program to the schools' specification? This is an opportunity for schools to join together and hire someone who can create the program for all of them, sharing the cost. This is also an opportunity for schools to access businesses or universities who may be willing to help create the program in order to meet the needs of the school district. No need for the business or university to be local. Even discussions and negotiations can be held online using programs like Skype or FaceTime.
Imagine how the writing process would be able to be changed. Teachers could require drafting and writing more often, without the unrealistic burden of grading so many papers. Time available for the individual, or small group coaching sessions would increase. True coaching, teaching, and learning would take place and student writing assignments can increase. The point is that in this and many other instances, there are existing technologies that can used in education, and will be, sooner or later. Computerized scoring of written work is right around the corner, and in some places, it has already arrived. But WE are the educators and WE know that it is the interaction between teacher and student that makes all the difference. So whether it is the use of Big Data Analytics or some other emerging technology, we will be led if we don't lead.
Our rejection of the concept of computer based scoring of written work is in part, our lack of knowledge of the possibilities and may be our fear of it taking teachers away from students. But as with all technologies, they allow a way to increase teacher contact time with students and improve student performance as a result. There are schools that are successfully using technology to create blended learning opportunities in which the same number of teachers have more time with smaller numbers of students in each class, while maintaining the same overall class load. We must lead the conversation by knowing and understanding how the technology can improve the educational process, which is based on the most important relationship between teacher and student. In educating our communities, it is essential to begin with the intention of improving teacher and student contact time, not replacing it. We need to design the solution, not be given it. First steps are opening our minds to the possibilities.
Zadrozny, Peter and Kodali, Raghu.Big Data Analytics Using Splunk (2013). New York: Springer Science+Business Media