Let me begin with a confession: the “Coefficient of Group Cohesion” (CGC) is a pseudo-scientific term that I came up with in order to make a very simple concept sound more academic. However, I make no apologies for this. After all, if making up pseudo-scientific terms were a crime, most social scientists would very quickly find themselves in jail! In this article, I will try to explain what the “Coefficient of Group Cohesion” is and why I think language teachers need to be interested in it.
Like any teacher, I began to notice early on in my career that some of the classes I taught were more successful than others. I am not talking here about individual lessons, but rather about series of classes with a particular group of learners. This phenomenon was particularly noticeable in Japanese university classes where I was teaching the same material using the same lesson plan to different groups of students. In these situations, I inevitably found that one of the classes became my “favourite.” I also noticed that learning outcomes in my favourite classes were better than in the others, even if the level of the students had initially been lower. After teaching a group of classes where this difference was particularly pronounced, I began to wonder what it was about my favourite group that made the class so successful. After much observation and reflection, I came to the realisation that the key factor was not my relationship with the students, but their relationships with one another. My favourite class was successful because everyone knew and got on really well with everyone else. There were also a couple of “characters,” which always helps!
Anyway, I started thinking of ways in which it might be possible for me to measure the degree to which each class got along. I thought that if I could find a way of assessing how well a group was socialised, it would help me decide how much time and effort I would need to dedicate to that side of things during the course. The idea I came up with was the “Coefficient of Group Cohesion,” which was essentially a numerical representation of the “cohesiveness” of each group of learners.
To calculate the CGC, I first asked each student to draw a map of the classroom using circles to represent people. In each circle, I asked them to write a “point value” representing how well they knew the student that the circle represented. The explanation I gave them was as follows.
0 points: Don’t know their name; never spoken to them.
1 point: Know their name, but never spoken to them.
2 points: Spoken to them in class, but not outside.
3 points: Speak to them regularly outside the class.
When they had finished, I asked each student to calculate the total number of points on their chart. I then collected these scores, added them together, and wrote the total number on the board for everyone to see. After that, it was just a simple matter of working out the score as a percentage of the maximum possible. For example, if a class had 20 students, the maximum possible score for every student was 60 (3 points for every student) times 19 (no points for yourself, obviously), which equals 1,140 points. If the actual total of points given was 570, then the Coefficient of Group Cohesion was 50%. Simple!
Another way of working out the CGC might be to use a points system where negative scores are given for people a student doesn’t like. The scale might look something like this:
I know and like this person.
I don’t know this person, but I feel positively disposed to them based on their appearance.
I do not know this person, and I have no feelings about them either way.
I do not know this person, but I feel negatively disposed to them based on their appearance.
I know this person, and I do not like them.
This system might actually produce a more accurate result, but there are obvious ethical problems with asking students to say that they do not like other members of their class. If you chose to calculate the CGC in this way, you would have to make sure that scores were collected anonymously. A major benefit of the original system is that it contains no value judgements, just simple factual assessments of how well people know each other.
So, is it really worth going to all this trouble to work out a number? I believe that it is definitely worth taking a small amount of time (the whole process normally takes around ten minutes) to find out something that I believe to be one of the strongest indicators of the likelihood of success in any given class. If I find that the CGC for a class is particularly low, I like to take one lesson to allow students to socialise in their own language. I give them a basic question, such as “How do you feel about learning English?” and ask them to discuss their answer in Japanese with as many people as possible before the end of the lesson. Allowing them to socialise without the pressure or embarrassment of having to speak English greatly increases the chance of a successful learning outcome for everyone. I also like to play a name-remembering game so that everyone tries to learn everyone else’s name within a 90-minute period. Having done all this, it is interesting to measure the CGC again at the beginning of the next class, as it will (hopefully!) have gone up significantly.
Although I have never used the CGC for anything other than improving my own classes, I have often thought that the concept could provide the basis for some interesting research. For example, a group of teachers could be asked to identify their “favourite” or most successful class. Those teachers could then be asked to work out CGCs in all of their classes to see if the ones they identified as being the most successful were also the ones with the highest CGCs. If such a relationship were shown to exist, it would provide a strong argument for devoting more time to allowing the students to socialise in their own language.
Please let me know if you have any thoughts or comments on this topic.