The Control-Group Conundrum

In scientific research, it is common practice to conduct an experiment with multiple groups, in which one group acts as the “control” group and one or more groups are experimental, with each experimental group subject to some type of treatment, i.e. experiment. Medical research, for example, may involve a control group that receives a placebo (i.e. a medication designed to have no medicinal effect) while an experimental group receives a medication being used on a trial basis.

While a placebo can work in medical research, classroom research involving second-language learning involves conditions that make use of a placebo untenable. Consider, for example, a study on pronunciation involving two sections of a course we will call Oral English. The same teacher is in charge of both sections. Section A receives the current mode of instruction being offered in the course, which involves no planned explicit instruction on pronunciation. Section A is therefore designated the “control” group. In Section B, on the other hand, the teacher plans to begin each lesson with a specific “treatment”, a ten-minute pronunciation activity to get students primed for the main speaking activities planned for that day. Section B is thus designated as the “experimental” group. The teacher plans to administer a pronunciation pre-test and post test to both groups.

The teacher’s belief is that the experimental group will do significantly better on the post test as a result of the treatment, provided that the two groups are equally proficient at the beginning of the course. Assuming that the two groups have equal mean scores on the pre-test (that is, not significantly different from a statistical perspective), is the teacher’s belief reasonable?

Unfortunately not. The central problem is that the control group is not receiving instruction that can be safely assumed to have a placebo (i.e., non) effect. Any student taking a course in Oral English, if they do have an opportunity to actually speak in class, may experience some improvement in their pronunciation. (Indeed, the teacher would likely hope so!) Moreover, the teacher has no control over what the students in the control group do outside of class. Some students, for example may have jobs that provide them with a natural opportunity to practice their pronunciation, and such students might actually experience more improvement in their pronunciation than students in the experimental group. The same can be said for any control-group students involved in dating relationships with English native speakers (or fluent non-native speakers), or for those involved in an on-campus club. The performance of such students on the post-test can offset the effects of the treatment administered to Section B and lead the teacher to incorrectly assume that the treatment had no real effect on the students’ pronunciation, when in fact no conclusion can reasonably be made about the effectiveness of the treatment  because of the performance of the control group.

Like pronunciation, vocabulary and grammatical development (to mention just two skills) are beyond reasonable control of any teacher. From a control perspective, this makes one-group experimental studies, involving one teacher and one group of students, much easier to manage. While some researchers might complain about the lack of a comparison group, the use of a pre-test post-test design provides teachers with the opportunity to compare students’ proficiency in a designated target skill (e.g. pronunciation accuracy) both before and after the treatment administered while avoiding the headaches of not being able to “control” members of a control group. Additionally, to account for the possible effects of independent pronunciation development, the teacher can elicit information on the post-test about the number of hours students spent on average weekly using English outside of class (not using any such data, or test-results, for research purposes without each individual student’s consent).

Experimental research in education is not as “neat” as that in the sciences. People are not lab rats, nor can they be placed in test tubes; their behaviour is beyond reasonable control. This does not eliminate the need for such research, however, nor its potential value for teachers. It does require, however, that we seek ways to make classroom research manageable so that we can achieve valid and useful results.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s