"A series of hierarchical linear models indicated that discussion-based approaches were significantly related to spring performance..." (Langer, 2009). As stated by Langer, discussions are mostly related to good and arising performance in language. This indicates group discussions can be implemented in a class of students who can converse well in the target language. I used to teach university students for six months and witness that students in that level of education participated well in the ESL classroom discussion. A lot of ideas were exchanged among one another since most of my the students are able to speak well in English.
On the other hand, students in lower proficiency level are not advised to be taught using classroom discussions in ESL class. This is because they find it hard to say out what they think in the target language. As a result, the discussions become a more of a first-language discussion classroom. For example, I was given a poor English proficiency class this year in SMK Chung Hua. Whenever I bring out a topic to be discussed in the class, the students were absolutely interested to talk about the topic - in their mother tongue! This has hindered my aim of improving their English proficiency.
For good classes, as mentioned earlier, I agree with the usage of classroom discussion as a method ranging from debates to group forum. However, for poor classes, I normally try to improve their vocabularies in order for them to give out their ideas - using games! For example, in teaching lexical chunks like 'In my opinion.....' or 'Frankly speaking.....' in giving their opinions. This will subsequently boost their confidence in English communication skill.
Lastly, surveys are utterly needed to give us an idea of students' preference of certain matter. For example, we can give a set of questionnaire to identify the students' preference towards classroom discussion