There are two sides to the debate around differentiated instruction. The first holds that differentiation, at least as it is commonly understood and practised, is dead because the evidence in support of this long-promoted approach appears chronically lacking. The second view holds that differentiation is not dead and merely pining for the fjords. It is this second view that still seems to hold sway in teacher education institutions, for writers of teaching standards, and, no doubt, in many schools.
An interesting new study from Flanders in Belgium provides yet another largely null result for differentiation. It comprises two randomised controlled trials involving a total of 2407 students in 200 classes in 65 schools.
Interestingly, the students were in Grades 8 and 9 and were learning about financial literacy. This extends an evidence base that is most frequently drawn from early literacy and numeracy teaching. The method of instruction was also quite significant. Students were involved in a ‘serious game’ where learning materials were supplied to them via computers and where they had to complete their work in workbooks with minimal intervention from the teacher.
The researchers seem to be under the impression that this teaching method has been experimentally verified to be more effective than a traditional approach, but the relevance for Australia and our post Gonski 2.0 turn towards technology-facilitated personalised learning is striking.
There were four conditions involved in the trials. In the control condition, schools did not receive the materials at all. I’m not clear as to the extent to which they then taught any financial literacy. In the second condition, schools were given the materials to work through and students were randomly paired-up. The third condition was the same as the second condition except that students were matched with other students of similar maths ability (which apparently correlates closely with financial literacy in PISA data). Finally, in the fourth condition, students were paired as in the third condition but the different ability pairs were given different levels of instruction. In this case, the higher ability pairs were given the original materials with the medium and lower ability pairs given materials with progressively more additional guidance. Comparing conditions three and four therefore measures the effect of differentiation (and comparing conditions two and three also does if you consider ability matching a form of differentiation).
Overall, there were no significant effects for ability matching pairs or for differentiated materials. There appeared to be some advantage of the differentiated materials for non-native Dutch speakers. There also appeared to be some positive effects on long-term retention. However, this was measured by a second, delayed post-test in the form of a homework task. The authors suggest this finding should be treated with caution due to the nature of the task and the ‘selectivity and size of the sample’. I must be missing something because I cannot find the data this is referring to.
So there we have it – another largely null result. Of course, the true believers will not be convinced. They will find some reason to argue that this is not true differentiation or that it has not been enacted properly. They may have a point. However, how many times do we have to fail to find Bigfoot before we are convinced that Bigfoot does not exist?
The ball is firmly in the court of those promoting differentiation. It is they who need to demonstrate that the right kind of differentiation can be effective. Until then, why should any of us assume this?