Do You Need a Large Sample
Do You Need a Large Sample
The short answer
No. Action research aims at local improvement, not generalization.
Depth over breadth
A small class studied deeply with several data sources is more useful than a large class with one quiz score.
Typical sizes
Five to ten for a focus group, twenty to forty for a whole class, sixty to one hundred for a multi-section study.
Do You Need a Large Sample
The short answer is no.
The long answer explains why.
Generalization versus local improvement
A large sample matters when you want to generalize. If you want to say “this works for Grade 7 students nationwide”, you need many Grade 7 students across many schools.
Action research is not trying to generalize. It is trying to fix something in one setting. A small, deep study of one class is exactly what is needed.
Depth over breadth
A study of forty students that goes deep, with interviews, observation notes, weekly tests, journal entries, and reflection, is far more useful than a study of four hundred students that collects only a single quiz score.
The action researcher trades breadth for depth. She gets fewer participants. She learns much more about each one.
Sample sizes that work for action research
A typical action research study has a sample in this range:
- Five to ten students for a focus group study on a specific issue.
- Twenty to forty students for a whole-class study.
- Sixty to one hundred students for a multi-section study, when the teacher teaches several sections of the same grade.
These numbers are small by formal research standards. They are appropriate for the kind of question action research asks.
When a small sample is not enough
There are a few cases when even action research needs a slightly larger sample.
- When the effect you expect is small. Detecting a small change in a small group is hard. The change can hide inside ordinary variation.
- When you want to compare two groups. A study of ten students split into two groups of five is too small to draw any comparison. You want at least twenty per group.
- When you plan to use statistical tests. Some tests need a minimum number of participants to be valid.
In a B.Ed. action research project, these cases are rare. Most studies are fine with twenty to forty students.
Describe the sample honestly
A small sample is acceptable, but only when it is described carefully and its limits are stated. Say how many students you studied, the grade level, the section, the gender mix, any relevant language background, and any way the group is different from a typical class in your context. Then state plainly what your findings can and cannot tell anyone else.
The one-line answer
A large sample is not necessary in action research because the aim is local improvement, not generalization. A small, well-studied class can yield valid and useful evidence.
One more card that sharpens the trade-off in a phrase you can hand back on a paper.
