08.10.15Dylan Wiliam advises: Forget the Rubric; Use Work Samples Instead
Part of the Solution or Part of the Problem?
In their new book, Embedding Formative Assessment, Dylan Wiliam and Siobhan Leahy make a simple, interesting, and really useful suggestion. Instead of using a rubric to communicate expectations for a given piece of student writing, Wiliam and Leahy suggest using work samples, instead–ideally two or more.
“Rubrics,” they note, “rarely have the same meaning for students that they do for teachers.” This, I think, is really insightful. We describe the things a good writing assignment needs but the description is too abstract and students can’t really visualize what we mean or how to do it. For example a recently published rubric designed to help teachers ensure kids master Common Core expectations notes that “Style and Conventions” is one of the key performance standards in a narrative piece of writing. An “inadequate” essay “merely tells experiences, events, settings and/or characters.” A “proficient” one “uses words and phrases telling details and sensory language to convey a vivid picture of the experiences, events, settings and/or characters.” So the difference is that a better essay uses sensory language and is vivid. But do students know what vivid language looks like? How to know when they see it or, more importantly, how to create it?
Instead Wiliam and Leahy suggests looking at student work, ideally two or more examples, and looking at the how–how are words made vivid? How are sensory details not only included but made effective in their inclusion—surely the addition of various sundry sensory details is not sufficient to make an essay strong. In fact, notes the ghost of Hemingway, they may make it worse.
I was struck by this advice in part because the technique of Show Call has been such a game changer in our classrooms, allowing us to embrace the power of the cognitive work students analyzing their peers’ written work do–it is specific, rigorous, focused–among the best things we do. Having seen that, the idea that the starting point for an assignment could be public work sampling-rather than the abstract rubrics we often use without even considering why–is especially compelling. And, of course, if you love your rubric there’s no reason you couldn’t use both work samples and rubrics to guide students through an assignment.
Thanks for posting about this, Doug. Shortly after Siobhan Leahy and finished the proofs of the book (of which the full title is: Embedding Formative Assessment: Practical Techniques for K-12 Classrooms), we came across a research study that shows that giving students contrasting examples is more helpful for improving writing than just providing good examples:
Lin-Siegler, X., Shaenfield, D., & Elder, A. D. (2015). Contrasting case instruction can improve self-assessment of writing. Educational Technology Research and Development, 1-21. doi: 10.1007/s11423-015-9390-9
This study is consistent with a large body of research that shows that giving students “only justs” and “near misses” are the best way to help students understand quality. It is the struggle to see the differences (what Robert Bjork calls “desirable difficulties” in learning) that is most helpful for long-term learning.
Of course, as Doug notes, rubrics have their uses, but we are convinced that rubrics are best thought of as a highly-stylized representations of our understanding of quality, and should be shared with students (or even better, developed with them) only after extensive investigation of samples of work.
I agree! I’ve used rubrics extensively in my technology classes for years. And yet, students still turn in research assignments that miss so many of the important elements I am expecting of them. So, last year, after one group failed the assignment across the board, I met with them and asked what I could have done differently to help them do a better job. Once one students told me that he really needed an example of a very well done assignment to understand the expectations, the others all chimed in and agreed. I then provided them with an example, and let them do the assignment again. Every students achieved at least a B, with most earning an A.
I intend to still use rubrics, but will always provide examples for research assignments now. I’m trying to make sure all of my middle school students know how to research and write a good paper, with NO plagiarism, using the Extract the Fact method. Providing examples has achieved what no amount of lectures or rubrics could.
Thanks, Doug and Dylan. This post reminded me of something I observed during a visit to High Tech High in San Diego: https://mpsmissiontohightechhigh.wordpress.com/2014/12/14/models-vs-rubrics/
What are the benefits of using work samples instead of rubrics to communicate expectations for student writing?