Does Differentiated Instruction Work?

I just received an email reminding me of the controversial commentary from James R. Delisle posted last month on Education Week. The title of Delisle's article is "Differentiation Doesn't Work". In this blog, I am posing it as a question instead. To answer such query, however, is not an easy task. Differentiated instruction is very complex as it involves assessment, planning and flexibility. All of these tasks hinge on the qualities of the teacher. A teacher who understands where his or her students stand is a good teacher. A teacher who tailors his or her lessons to maximize student's engagement is a good teacher. A teacher who can recognize that something is not working and needs to be adjusted is a good teacher.

Carol Ann Tomlinson is one of the pioneers of differentiated instruction. The Harvard Education Letter had a piece on differentiated instruction several years ago in which some of Tomlinson's views were highlighted:
While she would never say that differentiating instruction “is a piece of cake,” Tomlinson believes the approach is a path to more expert teaching. Like someone asked to make a meal, Tomlinson says, “You could have dinner with butter on toast with an egg. But if you want to grow as a cook, you need to expand your ingredients list.” Her four “non-negotiables”—a high-quality curriculum with clear goals, the use of data to monitor and provide feedback on student learning, the ability to recognize when something isn’t jelling and modify it to fit the student, and the creation of an environment in which students are supported and challenged—she says, “are not about differentiation. They are about a good classroom. That is good teaching.”
Differentiated instruction seems synonymous to good teaching. From the same article, a concrete experience from a teacher who is trying to differentiate is also cited:
Suddenly, “tiering”—or varying the difficulty of work for students based on readiness—had a twist: Kids didn’t like it when a classmate’s paper looked a lot different or had more problems on it. As she tried flexible grouping, students who seemed to need extra support actually “got it,” while those expected to glide would struggle. As Hauser put it in her write-up, “I quickly discovered that my assumptions were not always accurate.”
And the above is just the first step in differentiated instruction, understanding where the students stand. There is such a thing as incorrect assessment, and as in the above instance, wrong impressions on students can be made.  Doing the other steps requires even more from the teacher. To plan and to be able to change in the last minute definitely needs a good mastery of the subject and its pedagogy.

To appreciate how much should go into implementing differentiated instruction, the following list from Langa and Yost is quite helpful:

Content (Materials & elements)
  1. Select a variety of books and resource materials for handling variety in reading levels
  2. Select specific areas of interest within the focus area
  3. Use Learning contracts with students
  4. Group students according to readiness levels or interest levels
  5. Reteach to small groups who need support or explanations; exempt those who have mastered the material
  6. Establish learning centers or stations
  7. Allow students to work alone or with peers.
Process (how students gain understanding of main ideas and information)
  1. Use tiered activities (a series of related tasks of varying complexity)
  2. Use learning contracts based on readiness, interests, or learning profile of student
  3. Use independent learning
  4. Use choice boards, flexible grouping, reading buddies, learning centers and peer teaching
Products (ways students will demonstrate their knowledge or understanding of a topic)
  1. Write a story or a poem
  2. Write a book report, a play, or perform a play
  3. Debate or investigate an issue
  4. Design a model or a game
  5. Create a mural or a song
  6. Compare or contrast
Designing an experiment to evaluate its effectiveness is also very challenging because of the myriad of factors and its immense dependence on the quality of the instructor on how differentiation is implemented. Thus, it is not an easy task to find from research dependable experiments that measure the effectiveness of differentiated instruction.

More importantly, without doubt, differentiated instruction demands quite a lot from a teacher. The following excerpt copied from a guest post on a blog by a kindergarten teacher typifies one reaction a teacher may have with regard to differentiated instruction:

One teacher who started on something small, differentiated homework, finds "The results do not support the use of a differentiated homework structure for the acquisition of biology content or mastery of concepts." In primary schools in the Netherlands, a similar observation has been found: "Results showed that differentiated instruction has no statistically significant effect on student mathematics achievement, which was against expectations." These studies perhaps do not invalidate the differentiated instruction approach because one can always make the excuse that it is a failure of implementation.

However, one strong criticism comes from the very first step differentiated instruction requires. The article from the Harvard Education Letter also mentions:
But critics say differentiated instruction encourages teachers to categorize students based on popularized notions that may not actually be accurate or helpful in making content more accessible.
Daniel Willingham explains this more eloquently (also taken from the Harvard Education Letter):
“A lot of the time when we talk about differentiating instruction there is an implicit theory about the mind and the idea that different kids learn in different ways—and not only that, but that we have a deep enough understanding that we can then categorize kids on that basis. We assume that matching a teaching approach that plays to a kid’s strengths is the best way to teach. Or, should we work to attack areas of weakness? And how do we know if a teacher has correctly identified a child’s strengths? Differentiation sounds great, but on what basis are we differentiating? What do we know about this kid—and how do we know it?” 

Lastly, Greg Ashman made the following graph from PISA and TALIS data.

Above copied from Greg Ashman's Your own personal PISA - what does the TALIS show us?

This is simply a correlation, and it is weak, but is negative, that is, countries that perform well in PISA do less differentiated instruction. Obviously, it is not a solid proof that differentiated instruction does not work. But it also shows that there is no proof that it does.

Differentiated instruction taxes a great deal from a teacher. With the planning and management of different tasks or activities occurring inside a classroom, as Greg Ashman points, there are opportunity costs. If a teacher fails to do more compelling tasks (or even trivial ones, like collecting homework and helping a child organize), there is a price to pay for a teacher who spends more time on walking around the room just trying to keep the class in order during differentiated instruction.