This chapter is about an article by Benjamin Bloom (he of Bloom’s taxonomy) called the 2 sigma problem. 2 sigma is about significant effect sizes in research. They put in a little box explaining the concepts underlying standard deviation and effect size, but having read it twice I have decided knowing that significant effect sizes were found in this research is enough.
The research had three groups of students
- A class of 30ish being taught conventionally
- A class of 30ish students being taught with ‘mastery learning’
- Tutoring – alone or maximum with a group of three
They say the average tutored student score was higher than 98% of the students in the conventional class (an effect size of two which is massive). Which to a large extent makes sense though I would say even that would depend a lot on the tutor.
Bloom was trying to find a way to change whole class teaching to capture some of this and he suggests ‘mastery learning’. He calls it a ‘feedback corrective’ process. You teach them something, then test if they know it. If they don’t get 90% you go back and work on it some more until they do. I assume this parallels the fact that in a tutorial you can see clearly if a student has grasped something or not and adapt what you do accordingly. So Bloom came up with factors you could change in a class that might reduce the gap between tutoring and what he sees as conventional teaching in a normal class. One of those factors is what he calls mastery learning.
He assumes that in a normal class (a) you only test intermittently and that this is the teachers only access to a true understanding of who has grasped what and (b) teachers only get answers from proactive (and so bright) students, so the class moves on at their pace leaving many behind.
Perhaps it used to be like that, but my experience of watching people teach has been that some teach to the fastest and some to the slowest and both have their problems.
In what he calls mastery learning you teach a bit, test it, reteach where needed and don’t move on till it is mastered. And the paper describes the process as :
… classes were helped to review and relearn the specific prerequisites they lacked… the teacher retaught the items that the majority of students had missed, small groups of students helped each other over items that had been missed, and the students reviewed items they were not sure about by referring to the designated pages in the instructional material. It was algebra and French so a language.
And they got an effect size midway between tutoring and not doing this – so very positive results. But it does feel a bit like teaching to the slowest. Everyone is repeating stuff in various ways until everyone can understand. I can see the learning outcomes would be better, but won’t it take a lot more time to get there ? Are the groups streamed ?
They don’t talk about other ways to achieve individual feedback, though all require a lot more time investment from the teacher. You don’t have to test, you can set homework to get a clear profile of where everyone in a class is at. At a more superficial level you should be calling on a range of learners, not just the ones who always put their hands up first. Things like getting them to check in pairs before feedback could help with spreading the knowledge out a bit more.
You could use flipped learning so they are supposed to have worked on the material in some way before they arrive then you test at the beginning and teach what they still need help with. Though there you still have the group moving in lockstep. Or deliver material before class, then have tasks ready for them to work on in class so you can help individuals as needed. Or have fully differentiated materials so they work independently and there are tasks to stretch the fast finishers.
So while I can see his idea of mastery learning is slightly more teacher hours friendly than individual tutoring I think it will only increase learning outcomes significantly with the input of a lot more teacher preparation time. Which makes it considerably less revolutionary than it first sounds as then you are back to time, money and what institutions will support.
Bloom, B. (1984). The 2 Sigma Problem: The Search For Methods Of Group Instruction As Effective As One-To -One Tuto Ring. Educational Researcher, 13(6), 4–16.
sallyhirst 