Each spring, as they craft an organizational plan for the next year, our school administrators ask us what we value about our school. Whenever this happens I take it as an opportunity to survey my colleagues with a single question. When I first asked this question I was genuinely pondering the answer, but the unanimity of responses surprised me. Ever since I’ve asked to see if I still get the same responses, and I usually do.
The question I ask is: “Would you choose a larger single grade class or a smaller split grade class?”. I rarely find a teacher who prefers a smaller multi-grade class, with most preferring the larger single grade class. This may indicate how much teachers generally dislike split-grade classes, but I think it mostly demonstrates that the question of class size is more complicated that it seems.
The generally accepted wisdom is that smaller class sizes lead to better learning. Small class sizes are touted by independent schools as a way to attract students and they are one of the main ways they ways they make themselves more desirable to prospective parents and students. Recent research by Northwestern University supports this.
Professor Diane Whitmore Schanzenbach discovered that:
- Increasing class size reduces learning
- Larger classes also reduce the formation of “human capital” (perseverance, empathy, etc.) in students which leads to higher long term social costs
- Changes in class size affect at-risk students (e.g. low-income and minority) the most
These findings are counter to those of New Zealand education professor John Hattie who revealed in his book “Visible Learning” that class size doesn’t matter. Of course, it’s not quite that simple. What Hattie’s research showed was that on a list of factors that effect student learning class size mattered the least.
So how can we resolve these two apparently contradictory positions on the value of class size? I came across a possible explanation in Malcolm Gladwell’s latest book “David and Goliath“.
In “David and Goliath” Gladwell suggests that often, the things we think of as obstacles may, in fact, be our greatest advantages. Gladwell proposes an alternate model for how we think about class size to illustrate that position and inadvertently shows how both Hattie and Schanzenbach may be simultaneously correct.
When Gladwell talked with teachers about class size he found there was general agreement that large class sizes can impair learning. However he also discovered that teachers believe that the same is true about small class sizes. Teachers agreed that when classes became too small the group dynamics in a class became difficult, and individual students were more easily able to dominate the group and disrupt learning.
Based on this Gladwell suggests that the relationship between class size and achievement is actually not linear (as class size goes down learning goes up), but is best represented by an inverted U curve. As class size is reduced, learning improves until the optimum class size is reached. However if class size drops below the optimum learning declines.
I further believe that the optimum class size, where learning is maximized, isn’t a fixed number but a range. The optimum class size will vary according to the make up of the class and the various learning needs of students. The optimal class size for a high needs class will be lower than for a class of well-adjusted, independent students.
Consequently, the current “hard cap” system used in many schools is ineffective. In some schools a class of twenty-five primary students can learn effectively, and in others that number should be below twenty. Selecting a fixed cap number is inefficient in one school and reduces learning in another.
A system of class size ranges and school averages would be a more efficient and effective model. The educators who know the students should have greater flexibility and be empowered to make effective decisions that maximize learning.
After all, who should be making decisions that can affect learning so profoundly. The educators on site or someone in an administrative office manipulating school organizational models on a computer?