Ever heard of Bloom’s 2 Sigma Problem? Yeah, not the catchiest title is it? But it is a very interesting problem nonetheless. The problem is described by educational researcher, Benjamin S. Bloom, in his 1984 paper titled, The 2 Sigma Problem: The Search for Methods of Group Instruction as Effective as One-to-One Tutoring. The paper and references can be found here.

So what is the problem? Bloom’s research shows that one-on-one tutoring is more than 2 sigma - that is, 2 standard deviations - more effective than a conventional class (i.e. one teacher, many students). Don’t get overwhelmed if you don’t know what a standard deviation is. It is quite simple: what this means is that students who were tutored one-on-one scored, on average, 98% (2+ sigma) better than students in a conventional class. This diagram and example below might help. If there were a maths test out of 100 marks, this would be how the two groups of students scored.

Bloom's 2-Sigma Problem

So hopefully right now you are asking, why is that a problem? First, let’s take a step back for a minute and realise that the above data is factual. It stems from an experiment that has been performed many times and the result is always the same.

The reality is that in South Africa (and most the world) the average ratio of learner to teacher is 30:1. So the problem here is, as Bloom states, we “can’t find a method of group instruction (i.e. one teacher to 30 learners) that is as effective as one-to-one tutoring”. Simply put, one-on-one tutoring, although more effective, is impractical....maybe

Interjection: There is something I want to say here. Notice that Bloom, in his problem statement, is immediately narrowing down his solution to “Searching for a Method of Group Instruction as effective as one-on-one tutoring”. My interjection is simply to say that although group instruction is seemingly the only practical and economical (in both time and cost) method of instruction, we shouldn’t narrow our search just yet to only a method of group instruction.

Needless to say, Bloom didn’t find another method as effective (and to my knowledge, the problem has yet to be solved). The study goes on to look at other variables which improve a student’s achievement which are also beneficial in their own right. To name but a few these variables include; reinforcement, time-on-a-task, assigning homework, feedback loops, peer and cross-age remedial tutoring, social-economic status and change of curriculum. Looking at all of these is for another time.

It is also worth noting that when this study was done (30+ years ago) the level of technology integration in the world was far from where it is today. However the majority of new technology in the field (MOOCs, Google, Google classroom, Khan Academy, Moodle, etc.) aims to achieve what these studies were manually emulating. Clearly technology has a massive role to play in solving the problem of education, but more importantly we should be looking at what we want this technology to achieve.

This piece of writing, previous and future articles are all on a quest to find a practical, data-driven solution, to solve the education crisis in South Africa and the world. Note that I use the word practical, because action has to follow. I love discussing pipe dreams, but what is the pragmatic approach to those dreams? I also use the word solve, because I mean solve. You can’t solve a problem without a measurable goal to achieve. Education systems should always grow and improve, but there is a line somewhere that we need to cross to consider this "out of a critical state”: solved.

Completely Made-up Data for Emphasis

The reality is we need a goal and then some data before we can even begin to tackle this problem. We need to understand the problem. Bloom’s 2-sigma problems gives us data driven insight into what some possible solutions are.

Share This Post