Science Exams Don't Test Science

SATs exams are routinely used by many schools as standardised questions for class use and homework. After setting the school mandated homeworks for 13-year-olds, containing nothing but past exam questions, for their science homework, I have often been disappointed with the supplied marking schemes. The questions themselves are intended as summative test items to sample pupils' knowledge based on the 3-year long curriculum. What I needed for proper teaching was formative tasks based on what I had intended them to learn.

But issue the homeworks I did. And then the marking became a problem, not because is was onerous (there are few tasks that need less thought than marking test questions to a detailed mark scheme) but because the required answers were often incorrect or incomplete. Questions are written to correspond to specific curriculum learning targets, not in itself a problem, but when those targets are simplistic or read naively by the exam authors then science can go out of the window. Weak pupils gain credit for wrong answers because the question was not specific enough, and bright pupils lose marks because their perceptive answers went beyond the curriculum statements. The examiners often mistake a list of examples in the statements as being the limit of possible answers: for example, contributions to global warming may include the carbon dioxide from fossil fuel burning (obviously), but not the energy released from the same processes (a smaller, but real, contribution).

Similar problems have been caused by the current fashion for schools to purchase the exam boards' authorised textbooks, even though they are written to the test and encourage surface learning without depth. The worst problem by far for these texts, though, is the large number of errors in them and subsequent teacher responses. The errors are understandable given the short timescale for major changes imposed on the boards by the controlling authority, the Qualifications and Curriculum Authority (QCA). However, most colleagues I have discussed this with are unconcerned, with the majority seemingly happy to go with the flow. I have even been told by one head of department that we ought to teach what was in the book even if it was wrong. The reasons? Whatever was in the book will be marked as correct in any exam since the book was authorised by the exam board, and it is better to avoid causing confusion in the pupils!

The whole rationale for education has been subverted by the focus on exam marks. Exam marks are more important to students than knowledge. Subjects and exam boards are chosen on the basis of how lax their marking is to improve the students' chances and any attempt at rigour is seen as undermining the school's purpose.

The Sunday Telegraph has obtained documents from the QCA under the Freedom of Information Act:

Internal documents show that concerns raised by experts about accepting wrong answers in the test, taken by thousands of 14-year-olds in May, were overruled by the Qualifications and Curriculum Authority.

In a question which asked what organs a riding hat protects, the answer "skull" was accepted as correct - even though the skull is not an organ. Examiners were also told to award a mark to "ears" despite a graphic which accompanied the question clearly showing the riders ears outside the hat.

In another question, which asks pupils to describe how chalk changes when shaken in a container with granite, the word "weathered" was accepted as correct, against the advice of experts who told QCA that is was "completely incorrect".

QCA has known of the problems, but thought that correcting its exams would reduce the grade statistics for that year. Instead, teachers can continue to teach to the test , safe in the knowledge that a real examination of their charges' understanding will not be made until they get to university.