Some universities are using a controversial streaming system to bring weaker students up to speed. Are they getting results or are they simply reinforcing feelings of failure? Harriet Swain reports.

Agnetha Meadows was not surprised when she received a D in her maths A level but she was worried about how she would cope at university. Maths is a key part of her Leeds University degree course in the physics of electronics and instrumentation. "I like maths but I find it hard," she says.

In the end, she found the leap from school to university not as great as she had thought. This is because Leeds teaches maths to physics students in two streams. Agnetha, who was in the middle stream at her secondary school, is in the bottom stream at university, being taught concepts such as second-order differential equations by a former school teacher. The main difference from school is that students are given notes rather than having to make them themselves, she says. She has passed her first exam and her only worry now is what it will be like next year when the streaming stops. With streaming, she says, "it feels like it goes at more my speed".

Streaming, or "setting" to be more accurate since students are divided only in one aspect of their course, has been hugely controversial in schools. A review of evidence by the Scottish Council for Research in Education late last year suggested that it reinforced social and gender divisions. It found girls were generally placed in top classes and boys in lower groups, reinforcing peer pressures and stereotypes, while pupils from lower socioeconomic backgrounds tended to be overrepresented in the lower-ability groups.

ÖйúAƬ

Other criticism of the practice has focused on the difficulties of assessing pupils to decide which set to place them in. The argument is that setting on the basis of test results is too vague - are you measuring ability or mem-ory, swotting and workspeed? On the other hand, deciding on looser criteria is subject to the prejudices of teachers, who may be inclined to place well-behaved, middle-class children in higher sets.

The growing number of advocates of setting in ÖйúAƬ, however, claim that it is quite different at university level because what is being tested is not ability but how "well prepared" students are.

ÖйúAƬ

Sir Howard Newby, chief executive of the ÖйúAƬ Funding Council for England, speaking to the House of Commons Public Accounts Committee, recently blamed the new modular A levels for creating gaps in students'

knowledge when they arrived at university. His comments were part of evidence on a report by the National Audit Office on student achievement, which said that taking more mature students without formal qualifications might also lead to skills gaps.

Members of the PAC suggested that government policy to increase student numbers was another reason for the problem, because it put pressure on universities to drop entry standards.

Whatever the reasons, Michael Savage, an applied mathematician based in the department of physics and astronomy at Leeds, says there is now a huge spread of abilities in maths even among students who all have grade B at A level. Since Leeds admits physics students with Cs in A-level maths, and a few with Ds, teaching them all together has become impossible. Robert Curtis, head of maths at Birmingham University, says any university that does not stream is being "dishonest".

"People are coming into maths-dependent courses with a very broad range of backgrounds and abilities, so if you teach them all in one lump, it isn't really fair to the more able or the weaker ones," he says.

Birmingham divides first-year physicists into two groups according to A-level results and will soon introduce a similar system for engineers. At Leeds, physicists are divided into two groups following a "diagnostic test" on entry. Within the 40 per cent of students who make up the "less well-prepared" group, a further group of students is identified as most likely to drop out or fail. A part-time member of staff at the Open University takes them for an extra hour a week and may set up further special sessions, if needed.

The University of Manchester Institute of Science and Technology has been streaming students for six years in the maths modules of other subjects - mainly engineering but also chemistry, optometry and textiles. The maths department initially assesses students through a diagnostic test - students taking more mathematical subjects take a harder test - and combines this with previous test scores and a general appraisal of their abilities. Students are put in one of three groups for their first year-and-a-half. The groups used to be called high, middle and low but are now labelled with letters of the alphabet to avoid difficult connotations.

Colin Steele, director of service teaching at Umist, says the system does cause problems when comparing students at the end of the year - should students who have taken harder courses receive extra credit, for example? - but overall he is sure it works. Fewer weaker students are dropping out or failing to keep up, while there appear to be benefits for brighter students too.

ÖйúAƬ

Other approaches, such as putting on extra classes for weaker students, did not prove so successful because they did not feed into assessments and take-up was poor, Savage says.

ÖйúAƬ

While streaming or setting is expensive because it involves employing an extra lecturer, the evidence so far is that it is more than cost-effective. Preventing just one or two students from dropping out pays for itself. "It breaks down the culture of hostility towards maths and stops students thinking they are no good," he says. "It builds competence and boosts confidence. But you need people to teach it as a school teacher would teach it."

At Leeds, Tom Roper, who takes the "less prepared" stream, is a former school head of maths. At Birmingham, PhD students often take this stream. But this is where problems can arise, warns Judith Ireson, reader in psychology and education at the Institute of Education, in London. Ireson, whose book Ability Grouping in Education was published in September, says:

"There is a tendency for teachers to change the way they teach pupils in the lower sets so those pupils do not get the same quality of teaching. That is something to watch out for."

Her research, which involved school children, concluded that setting could have limited advantages in maths - it was possible to boost attainment for the higher-attaining pupils - but there was a tendency for pupils in the lower groups not to do so well. She says previous research has suggested that pupils in lower streams tend to become disaffected and demotivated and miss the stimulation of mixing with brighter pupils.

Not all university maths departments are convinced either. Peter Avery, head of maths and statistics at Newcastle University, says the department did once pick out those students with poor maths A-level grades and put them into special tutorials, but without much success. "I have always felt the important thing is how much work a student is willing to put in," he says. "If they have problems, so long as they put in the work and ask questions they will do fine." He concedes that the range of ability is now wider than it was. "The weaker end just has no concept of numbers or even what a form-ula is or what brackets mean," he says. "We get told by the departments that they want us to teach them quite complex statistical methods and they cannot do simple manipulation. It's very difficult."

But he still believes the main problem is motivation rather than ability. Those who are weakest on entry do not necessarily end up the weakest.

This is something that the University of Southampton has tried to accommodate in a complicated system of setting used not in maths but in languages. It has introduced a programme of seven stages, with stage one for beginners and stage seven for advanced learners. Students are expected to progress by at least one stage a year and must reach stage six or above to graduate with a language named in their degree. Reaching stage five in one language will allow them to graduate with a degree in contemporary Europe but they will have to present transcripts of marks to anyone wanting details of their abilities in a specific language.

The system allows people to advance more quickly if, for example, their skills rapidly improve after a year abroad or they find a language easy because they already know one that is similar. It is also helpful in meeting increasing financial demands to teach students studying the same language together, even if their degree programmes are very different - French with engineering and French with Spanish, say.

Michael Kelly, professor of French and director of the subject centre for languages, linguistics and area studies at Southampton, says students' modern-languages abilities when they arrive at university are becoming more varied.

"We take the view that it is not reasonable to expect these students to start and finish at the same level," he says. "We have noticed the benefits of realistic expectations. It gives them achievable aims, whereas if you have too wide a disparity in levels of language attainment, some students get bored while others get demoralised because they cannot keep up."

For Kelly, the changing face of the student population, with more part-time, mature and overseas students, means that the Southampton scheme, which has already sparked interest from other universities, is the way ahead. He says: "Once you start loosening the frontiers between different sectors of education these issues are going to affect all disciplines."

ÖйúAƬ