School comparisons miss too many factors to be effective, says Mervyn Flecknoe
Chris Woodhead, the chief inspector of schools, has just released Performance and Assessment Reports (Pandas) to 24,000 schools in England. These will show how each school compares in results with others of similar intake and characteristics. They will become the next set of data by which heads and governors will be judged.
The idea of comparing the results of schools which have a similar intake is seductive and Pandas are here for a while at least. They are here until people realise what a waste of effort and a damaging use of emotional energy they will be.
The studies on school effectiveness all show that differences in easily measurable social characteristics such as free school meals and prior attainment explain some of the differences in the performance of children.
More detailed studies identify up to eight less easily-measured features of family life which make a difference. If all of the differences could be explained by external social factors which we could measure then we would say that all schools were having the same effect and that we had an education system of acknowledged uniformity. As it is, when all indicators have been taken into account, there is still unpredictability, there is a scatter.
Mr Woodhead implies that whatever is unmeasurable is under the control of schools, and that if he publishes the variation, schools will improve because they will begin to affect the unmeasurables. These assumptions are without logical foundation and will lower the morale of the teachers who deal with the most intractable educational problems. Teacher morale is an indicator of children's achievement.
One of the excellent instruments available to secondary schools to help them to assess their performance is called Yellis (Year 11 Information System). A series of tests, administered at the beginning of Year 10 is used as a baseline. Children are placed in one of four bands. The group in the fourth band, the lowest achievers on the test, have about a 2 per cent chance of achieving a grade B at GCSE, whereas the modal performance of the band is grade F. So why do some children with such low prior attainment score so well at GCSE? We have absolutely no data to explain this.
The truth of the matter is that a whole range of factors may play a part. Some of these factors may be under the control of the school, it is likely that more are not.
Here are just two examples of factors which are not easily measurable but which do influence the achievement of children. Music seems to be important. The Portway secondary school in Bristol, where every child learns to play a brass instrument on entry, over-achieves. Why does music make a difference? It could be that charismatic music teachers are "significant adults". Contact with a significant adult is one of the proven factors which make a difference. In today's financial climate music teachers have to be a bit different to survive and many rely on intense loyalty from their performers. These performers do better than expected at GCSE.
We know that city technology colleges sometimes have high added value. Discounting the special funding arrangements, they pick their parents. Children who receive free meals but who have supportive parents are more successful than children with wealthy parents who let them go to night-clubs instead of doing their homework or getting their beauty sleep. They enjoy strong parenting even though they do not appreciate it at the time. "You will do your homework before you watch the television" is a recipe for success in music and GCSE examinations.
There is no way at present of measuring parental support for education. The idea of depending on data which only concerns the easily measurable was discredited by Robert McNamara, who was Secretary of Defence in the USA when the Vietnam war was going so badly. He found out that his military were using body counts as the only indicator of success. They had tried to find other ways of showing that they were winning but they were difficult to use.
You may remember the My Lai massacre which boosted body counts a lot but made no difference to the progress of the war except to make the North more intransigent and determined. He said that they had to make the important measurable and not simply make the measurable important. This is called the McNamara fallacy.
There are three areas of ignorance which affect even Mr Woodhead: 1. We do not yet know in a systematic way what are the process indicators of successful schooling. Processes are the aspects of schooling which are under the control of teachers and governors. There is no agreement about which processes are most productive. Schools can be held accountable for their processes and it is these which thoughtful parents use to make judgments about schools. A common one is the health of the orchestra.
2. We cannot measure some of the outputs which we consider so valuable, for instance, capacities for aesthetic appreciation, development of self direction in learning, mathematical creativity, persistence and realistic self-concept. These are not necessarily measured by GCSE results.
3. We cannot measure some of the input measures which are important in deciding which children will do well and which will under-perform.
To compare schools which have "similar intakes", based only on measurable factors, and then to beat the less successful school with the comparison, is a crass act of uninformed and unintelligent cruelty which has less chance of raising standards than most which have been thought up so far. Parental choice will soon reinforce any differences found.
What is needed to raise standards is a concerted effort to turn schools into learning organisations, with rewards and tax breaks for the teachers who research their own practice; to give headteachers and governors proper powers to maintain teaching standards in conjunction with unions which take on board the necessity to maintain standards of professionalism and to campaign for better funding for the least advantaged in our society for the benefit of all.
What we need are not Pandas but something with more thought and sensitivity. We need to realise that not everything in black and white makes sense.
Mervyn Flecknoe was headteacher at a comprehensive for 14 years, and now lectures in the school of professional education and development at Leeds Metropolitan University.