# understanding research

effect size

WHAT?

An effect size is how much of an effect something has on something else. The larger the effect size, the greater potential impact.

EXAMPLE

A series of research studies are carried out. Learners are given the same instruction. Each study splits a class of learners into 2. Half the learners are given instruction verbally and subsequently make written notes on the content (the control group). The other half are given the same content but asked to self explain the content before they make written notes (the experimental group). A week later, all learners are given a test where they demonstrate their knowledge and understanding. The effect size is how much of an effect self explaining has on the performance of the learners.

COHEN'S D

Using Cohen's d is one of several methods for measuring effect size. It is commonly used in educational research and illustrated in the images below, measuring the distance between the 2 peaks. Imagine the bell curve below is from a standardised test where the light blue line is the control group and the dark blue the experimental group. The first image shows Cohen's d (and therefore the effect size) is 0.2, a small effect size. The second image shows where this becomes a medium effect size (0.5) and the third image shows where this becomes a large effect size (0.8). In reality, an effect size of 0.4 is considered statistically significant, meaning it's worth considering adopting.

## boundary conditions

EXPLANATION

In educational terms, boundary conditions are the environment (conditions) that a theory must operate within (boundaries) for it to have the expected impact. Put simply, the who, where and when that must be followed for a theory to impact as expected. Stepping out of these boundaries could lead to a dimmer-switch like decreasing impact, or an instant on-off-switch like effect where there is no longer an impact at all.

WHO

In educational terms, boundary conditions are the environment (conditions) that a theory must operate within (boundaries) for it to have the expected impact. Put simply, the who, where and when that must be followed for a theory to impact as expected. Stepping out of these boundaries could lead to a dimmer-switch like decreasing impact, or an instant on-off-switch like effect where there is no longer an impact at all.

WHERE

Examples: Groups and pairs: Impact may only occur when learners work individually. Instruction: This is often the variable in research but can also be a boundary condition, for example: When instruction does not include a summary first, impact may be lost.