Stay up to date with notifications from The Independent

Notifications can be managed in browser preferences.

Poor marking flaws primary tests, says study

Judith Judd
Thursday 06 March 1997 00:02 GMT
Comments

Serious flaws in national tests for 11-year-olds, which will form the basis of next week's government league tables for primary schools, are revealed in a study published today.

The research from King's College, London, says that the questions and marking of the English tests were so bad that nearly one in four children was given the wrong "level" or grade. Most received a lower level than they deserved.

Gillian Shephard, the Secretary of State for Education, originally planned to wait until the tests had "bedded down" before publishing primary league tables but was persuaded by Michael Heseltine, the Deputy Prime Minister, to change her mind. Labour said last week that its new literacy targets would be based on the tests.

Government exam advisers who supervised the tests immediately condemned the study as "flimsy".

In science, the researchers found, 1 in 10 children was given the wrong level and some questions were so badly worded that they could be answered by guesswork or without any knowledge of science.

Researchers re-marked 338 test scripts by 143 pupils in 10 schools. The expected "level" for an 11-year-old is 4. Pupils graded Level 5 are two years ahead of their age.

In English, marking mistakes were found in every single paper, in 84 per cent of science papers and 54 per cent of maths scripts. The level changes mainly involved more able pupils, who moved from Level 4 to Level 5.

In story writing, the researchers argue, the markers were prejudiced against topics such as football, adventure and crime, often written by boys, and in favour of twee, "Disneyesque styles characterised by 'little' and 'lovely'."

Nearly two-thirds of the questions in the English reading test were unclear, the report, commissioned by the Association of Teachers and Lecturers, said.

One question asked pupils to say whether or not a boy got on with his brother and to explain their reasons for choosing yes or no. There was no indication that an extra mark was awarded only if they said that he both got on with him and did not get on with him.

In science, says the report, "correct answers could be achieved by a lucky guess". In a question on magnets, for instance, candidates had to choose between the words "repel" and "attract" so there was a 50 per cent chance of obtaining each answer without knowing anything about magnets or the meaning of either word.

Professor Margaret Brown, professor of maths education at King's, says in the foreword: "The study shows that both the tests and the marking have not yet reached sufficiently high standards to justify teacher and public confidence. Its findings are important because of the major consequences of these tests for schools, teachers and pupils."

Dr Nick Tate, chief executive of the School Curriculum and Assessment Authority, called the research flimsy and biased. The authority's own study by Bath University of 10,000 scripts from 1,600 pupils had found none of the flaws.

Children's stories

"Frankie was a little grey mouse of about two years old. His mother had died when he was only three weeks old leaving him all alone in the world."

And later: " 'Yummy, yummy, yummy in my tummy,' said Frankie as he edged nearer to the great slab of cheese."

Examples from a "twee" story awarded a high Level 5 which researchers say was only worth Level 4.

"Kevin was getting really frustrated because it just wasn't his day. He kept missing the ball and was always getting tackled when he had the ball. It was his worst performance on the football pitch ever. The worst thing was that he had been dropped from the five-a-side tournament because of his play in recent days."

Example from a story awarded Level 4 which researchers say deserved Level 5.

Join our commenting forum

Join thought-provoking conversations, follow other Independent readers and see their replies

Comments

Thank you for registering

Please refresh the page or navigate to another page on the site to be automatically logged inPlease refresh your browser to be logged in