Wednesday, 3 July 2013

The fate of bright children: Streams, sets, and educational delusions

Education ought to be a straightforward business. You teach the kids as much as possible of all they need to know. Sure, you have to work out roughly what they need to know, but reading, writing and arithmetic are the core, because if they can do that they can work out a lot of the rest for themselves. Also, you need to work out the best way to teach them. Theories abound, but some techniques and processes can be supported by evidence. Tasks must be graded from easy to difficult, and when children find tasks difficult they must be helped to understand them, usually by an explanation which involves going back to simpler steps and building upwards again to more complicated constructions.

Basically, all educational systems have an inherent understanding that knowledge varies in intellectual complexity, and that some simple tasks have to be mastered before more complicated tasks can be attempted.

The Royal Society is simply the Royal Society. It is not the Royal Society of Science (but it is) or the Royal Society of Thinking (but it is) but just the RS and its Fellows are FRS. It has many merits, but two stand out: the main lecture hall has good acoustics, and they serve tolerable coffee at half time. As to content, although it varies, it generally oscillates around a very high standard. Their public lectures are entertaining and their conferences some of the best days out that London can provide. FRS is the highest accolade in British science, and the holders of this cherished award are excellent on their own subjects, and probably well above the average on all subjects, but with some inevitable regression to the mean.

So, it was with great pleasure that I joined a conversation between two FRSs over coffee and chanced to hear this exchange: “We know the difference between people in terms of physical strength, but what is the difference between people in intellectual power”. Seizing my opportunity I leant forwards and said “Fivefold”. They seemed satisfied by this and continued their conversation.

Jensen gave, as a rule of thumb, the observation that if any intellectual task is set to a class of children it will be found that one child can do it five times faster than the slowest child. The difference is not as obvious as getting them to run a race, but it is obvious to any teacher.

How should teachers cope with these disparities in intellectual power? The most common approach seems to be to pretend that it is not there. However, every now and then a comparison comes to light between those who do not recognise cognitive differences and those who do. Many schools teach children in “mixed ability” classes, and do only some teaching in sets or streams. Worse, teachers assume that teaching to the middle range will be good enough, and that bright children do not require any extra attention. Indeed, in such schools teachers often do not know who the bright children are.

Ofsted, the office of raising standards and improving lives, is in fact a government department which casts an uncomprehending eye on government schools. In June they emitted a confused report entitled: The most able students: Are they doing as well as they should in our non-selective secondary schools?

Their answer was no. This is not a surprise. Non-selective schools put themselves at a disadvantage, unless they put children in different learning streams, because a fivefold power difference in intellect makes children hard to teach en masse.Slower learners struggle, faster learners struggle with boredom. Gottfredson’s summary of the US Army training data shows just how different the learning trajectories are for different levels of intellect.

So, I should be happy that this report recognises, without being too clear about it, that teaching children of very difference intellects in the same classroom is unwise.

However, much as I want to welcome the report, I am not sure that it meets the minimal standards for a publication. The authors do not mention regression to the mean, correlations between variables, or reliability. Kids who do well at 11 on one set of tests are not guaranteed the same level of excellence on later exams years later. The reprt is bereft of any technical appendix. I will see if one is available before commenting in detail. The comparison between the outcomes achieved by comprehensive schools and private schools (the latter do much better) is not all that fair, in that the children and the families very probably differ.

So, it is good to see intellectual differences recognised, but a pity it has been done in such an uncomprehending way.


  1. "Indeed, in such schools teachers often do not know who the bright children are."

    That is a very interesting statement to me. Is there research saying this?

    Neither in my own schooling nor in my children's have I noticed this sort of ignorance. There were and are plenty of teachers who react to bright children with ill-concealed hatred. This emotional reaction leads some of them to call brilliance by other names, but how do some of them fail to notice brilliance? Gifted kids are obviously different, except for the ones who have learned concealment.

  2. From the report, page 7 "The survey findings present a discouraging picture of what it means to be one of the most able students in non-selective secondary schools in England. The 2,327 lesson observation evidence forms scrutinised separately as part of this report showed that the most able students in only a fifth of these lessons were supported well or better in such schools. Moreover, in around 40% of the schools visited in the survey, the most able students were not making the progress of which they were capable. In a few of the schools visited, teachers did not even know who the most able students were." So, a minority of schools, but it is very surprising nonetheless.

  3. In every human activity that I know of, some people are better at it than others.
    In every human activity that I know of, specialisation works.

    Therefore schools should be streamed. In big cities the most effective way of streaming is probably to have different schools for different ability groups; in small places, streams within the same school.

    I used to have a colleague who preached the virtues of unstreamed comprehensive schooling. But when he gave "supervisions" at Cambridge, teaching students in groups of two, he always paired them as two Firsts, two Upper Seconds, and so on, using their results from previous years. Perhaps he thought that other people wouldn't notice.