The success of any classroom intervention is to a large part dependent on the teacher. The teacher has the power to create an environment which will facilitate metacognition or one which will not. However, any teacher’s behaviour is infl uenced by her own knowledge, beliefs and opinions. It can be diffi cult to differentiate between these three concepts; opinion and beliefs infl uence knowledge and knowledge contains belief. Nespor (1987) differentiates between belief and knowledge in a number of ways. Firstly, beliefs are dependent on feelings and on evaluating knowledge, so that a teacher’s beliefs about the usefulness of a particular strategy or knowledge system will affect the way he teaches this area. Secondly, beliefs are thought to be stored in episodic memory and are drawn from previous emotionally charged experiences, whereas knowledge is thought to be stored semantically, in a more organised and objective manner. Thirdly, belief systems are not necessarily consistent, nor are they necessarily open to group consensus, whereas we would expect knowledge to have some kind of objective consensus and be accountable against evidence. Beliefs may be fi ckle, easily changed or long lasting. Opinion lies somewhere in between these two. Our opinions tend to be based on our value system, but they also usually make some claim to be rational or logical and draw on knowledge for support. Of course opinions are often biased in the way in which one selects aspects of a knowledge system and that selection can be driven by a strong belief. Possessing knowledge about something is no guarantee of how that knowledge will be demonstrated. Teachers possessing similar knowledge about a construct such as metacognition may act on this knowledge in very different ways, depending to some extent on how useful they believe this knowledge to be. Beliefs and opinions arise out of experience and all teachers were once students in classrooms. The experiences they had as students are likely to infl uence their beliefs and opinions about teaching; to affect their behaviour in their own classroom; and to infl uence the experiences they provide for their own students, some of whom will go on to become teachers. It is clear that any intervention into classroom practice must take account of the teacher’s beliefs, opinions and knowledge about the proposed teaching strategy. Many researchers have bemoaned the fact that their carefully theorised and planned classroom intervention study looked

completely different when implemented by different teachers across the sample. Changing well established classroom practice is a diffi cult and frightening prospect for many teachers, and researchers need to take account of the emotional effects of intervention studies on the teachers involved.