Nothing shows the fall in educational standards better than the failure of the current generation of journalists and teachers to understand the various types of examinations.
This article was originally written in 2012 but still applies. A postscript for the effects of COVID has been added.
GCSE grades have fallen this year (2012) for the first time.
The pupils, parents, schools, higher education industry and even the government have a vested interest in ensuring that grade inflation continues until everyone gets a grade A. It is only the long term health of industry and the country that demands proper examinations. So what went wrong?
Apparently this tiny change, that has set the educational establishment and media squealing like stuck pigs, is due to a tougher marking and standardisation regime. What shocks me about this whole affair is that teachers, media and government seem to have no idea about how to set exams. How and why should we set exams?
The government must decide whether it wants GCSEs to be a graduation certificate that shows that the student has learnt the basics of a subject or a test of ability. If the GCSE is to be used to test ability then it should be a "selection test". In a selection test 5% of students might get an A grade, 10% a B grade, 10% a C grade etc. and 40% fail. The test selects a proportion of the students, grades are given to fixed percentages of those taking the test so the same number of students get grade Cs every year. A selection test will involve difficult problems that can only be solved by clever students and it does not simply test how much irrelevant trivia they can remember. Selection tests are much fairer than the current GCSE because they assume that students are the same cleverness from one year to the next - at present it was much harder for a student to get a B Grade in 1995 than an A Grade in 2011 but employers will take the A Grade applicant without realising that this has occurred.
The A Levels were a selection test until the mid 1980s:
You can see this clearly from the way the proportion of A Grade passes is constant until the basis is changed. The GCSEs, which were introduced in the late 1980s, have never been a selection test.
Perhaps there should be two tests, a graduation certificate and a university entrance exam....
The recent GCSEs raise a question about how far teachers should be involved in assessment. Teachers are being involved more than ever in assessing their students and this is a dangerous and fallacious practice. Many teachers feel that they know all about their students but objective studies show that this is not the case (Eckert et al 2006, Bentz and Fuchs 1993, Kilday et al 2011, Begeny et al 2011, to mention but a few of a long line of studies that show teachers tend to assess behaviour and attitude rather than ability and attainment). Teachers do not just incorrectly assess ability but are also systematically biased, especially against boys, especially female teachers against boys (Mullola et al 2012). It can only be concluded that teachers should be kept as far away from any important assessments (those that affect a child's future) as possible. Continuous assessment (which is a brutal and direct discrimination against teenage boys) and other forms of assessment on the cheap are no substitute for external, direct assessment.
Postscript (January 2021):
This article was originally written in 2012 but still applies. A postscript for the effects of COVID has been added.
GCSE grades have fallen this year (2012) for the first time.
The pupils, parents, schools, higher education industry and even the government have a vested interest in ensuring that grade inflation continues until everyone gets a grade A. It is only the long term health of industry and the country that demands proper examinations. So what went wrong?
Apparently this tiny change, that has set the educational establishment and media squealing like stuck pigs, is due to a tougher marking and standardisation regime. What shocks me about this whole affair is that teachers, media and government seem to have no idea about how to set exams. How and why should we set exams?
The government must decide whether it wants GCSEs to be a graduation certificate that shows that the student has learnt the basics of a subject or a test of ability. If the GCSE is to be used to test ability then it should be a "selection test". In a selection test 5% of students might get an A grade, 10% a B grade, 10% a C grade etc. and 40% fail. The test selects a proportion of the students, grades are given to fixed percentages of those taking the test so the same number of students get grade Cs every year. A selection test will involve difficult problems that can only be solved by clever students and it does not simply test how much irrelevant trivia they can remember. Selection tests are much fairer than the current GCSE because they assume that students are the same cleverness from one year to the next - at present it was much harder for a student to get a B Grade in 1995 than an A Grade in 2011 but employers will take the A Grade applicant without realising that this has occurred.
The A Levels were a selection test until the mid 1980s:
You can see this clearly from the way the proportion of A Grade passes is constant until the basis is changed. The GCSEs, which were introduced in the late 1980s, have never been a selection test.
Perhaps there should be two tests, a graduation certificate and a university entrance exam....
The recent GCSEs raise a question about how far teachers should be involved in assessment. Teachers are being involved more than ever in assessing their students and this is a dangerous and fallacious practice. Many teachers feel that they know all about their students but objective studies show that this is not the case (Eckert et al 2006, Bentz and Fuchs 1993, Kilday et al 2011, Begeny et al 2011, to mention but a few of a long line of studies that show teachers tend to assess behaviour and attitude rather than ability and attainment). Teachers do not just incorrectly assess ability but are also systematically biased, especially against boys, especially female teachers against boys (Mullola et al 2012). It can only be concluded that teachers should be kept as far away from any important assessments (those that affect a child's future) as possible. Continuous assessment (which is a brutal and direct discrimination against teenage boys) and other forms of assessment on the cheap are no substitute for external, direct assessment.
Postscript (January 2021):
During the COVID epidemic in 2020/21 there have been calls for teachers to assess exam grades. The BIS commissioned "Investigating the Accuracy of Predicted A Level Grades as part of 2009 UCAS Admission Process" in 2011 (using 2009 data) that found that teachers over-predicted grades. In 2009 grade inflation had resulted in 33.11% of pupils obtaining an "A" grade which meant that simply predicting that all pupils would get an A had a 33% "accuracy".
The BIS study found that teacher predictions of "A" grades were superficially 63.75% accurate but the real figure needed to take account of the large excess of A grades. Teachers seemed well aware that grade inflation means that 61% of candidates would get an A or a B and so systematically biased their assessments to predict high grades. Their assessments would be correct 61% of the time by chance alone for predictions that pupils would get either A or B grades. This has an amusing consequence: if children were all awarded A or B grades this year it would be about as fair as awarding grades on the basis of teacher predictions.
The assessors can only be assessed accurately by looking at B, C and D grades where predicted grades can be both higher or lower than the attained grade. On B,C and D grades the teachers were c.40% accurate. Or to put it another way: the teachers were wrong 60% of the time. This is consistent with the research on teacher assessment. Had grade inflation not occurred the teachers would be hopelessly wrong in all their predictions.
There are only five grades and the difference between an "A" and a "B" is often the difference between Cambridge (Russell Group) and Keele (other) in outcome. So an error of one grade can damn the child.
It is probably the case that teacher assessments should be used when no other method is available but broadcasters etc. should not state that "teachers know their children well and so are best placed to assess them". The reason we spend millions each year on examinations is because teachers are not best placed to assess their pupils.
See also:
Record A Level passes again
Eckert. T.L et al (2006). Assessment of mathematics and reading performance: An examination of the correspondence between direct assessment of student performance and teacher report. Psychology in the Schools
Volume 43, Issue 3, pages 247–265, March 2006
Bentz. J.L. and Fuchs. L. (1993) Teacher Judgment of Student Mastery of Math Skills Assessment for Effective Intervention January 1, 1993 18: 219-232
Kilday et al. (2011) Accuracy of Teacher Judgments of Preschoolers’ Math Skills. Journal of Psychoeducational Assessment August 1, 2011
Begeny, J.C. 2011, Teacher Judgments of Students’ Reading Abilities Across a Continuum of Rating Methods and Achievement Measures School Psychology Review,
2011, Volume 40, No. 1, pp. 23–38
Sari Mullola, Niklas Ravaja, Jari Lipsanen, Saija Alatupa, Mirka Hintsanen, Markus Jokela and Liisa Keltikangas-Järvinen (2012). Gender differences in teachers’ perceptions of students’ temperament, educational competence, and teachability. British Journal of Educational Psychology Volume 82, Issue 2, June 2012, Pages: 185–206,
Written 25/8/2012
See also:
Record A Level passes again
Eckert. T.L et al (2006). Assessment of mathematics and reading performance: An examination of the correspondence between direct assessment of student performance and teacher report. Psychology in the Schools
Volume 43, Issue 3, pages 247–265, March 2006
Bentz. J.L. and Fuchs. L. (1993) Teacher Judgment of Student Mastery of Math Skills Assessment for Effective Intervention January 1, 1993 18: 219-232
Kilday et al. (2011) Accuracy of Teacher Judgments of Preschoolers’ Math Skills. Journal of Psychoeducational Assessment August 1, 2011
Begeny, J.C. 2011, Teacher Judgments of Students’ Reading Abilities Across a Continuum of Rating Methods and Achievement Measures School Psychology Review,
2011, Volume 40, No. 1, pp. 23–38
Sari Mullola, Niklas Ravaja, Jari Lipsanen, Saija Alatupa, Mirka Hintsanen, Markus Jokela and Liisa Keltikangas-Järvinen (2012). Gender differences in teachers’ perceptions of students’ temperament, educational competence, and teachability. British Journal of Educational Psychology Volume 82, Issue 2, June 2012, Pages: 185–206,
Written 25/8/2012
Comments