National IQs and PISA: it changes everything

Roger Titcombe's picture

This article shows that the East Asian education systems that come out at the top of the PISA rankings sink to the bottom half of the table when national IQs are taken into account.

If a given group of pupils in the same class all have 100 per cent attendance, with the same teacher and the same amount of teacher attention, would we expect them all to perform equally in the end of term exam? Of course not. We would expect the ‘brighter’ pupils to outperform the ‘less bright’ ones. Hundreds of English schools use Cognitive Ability Tests (CATs) taken in Y6 as the basis of admissions systems designed to produce ‘balanced comprehensive intakes’ in which the ability profile matches the national ability profile so far as possible. Part 4 of my book, ‘Learning Matters‘ explains how this works in the London Borough of Hackney and at Mossbourne Academy, one of the Hackney schools.

In my article about the government’s ‘Progress 8’ secondary school accountability system, of which I am highly critical, I argue for an alternative CATs – based system.

My evaluation of school effectiveness approach requires the production of scatter diagrams like the one shown in this article for Cumbria schools, where in the 1990s they were produced by the LEA every year, but largely ignored.

Now think about all the schools in England. If they were all equally effective and followed the same curriculum then they would all get equally good GCSE results right? Of course not, but here mainstream educational thinking takes a false turn. The almost universal assumption is that some pupils are disadvantaged by poverty, poor parenting or ethnic discrimination and that this accounts  for their poor performance. There is much talk of an ‘achievement gap’. In my book and in this article, I use evidence from my study of Mossbourne Academy and the Hackney admissions system to argue that this ‘achievement gap’ theory is incorrect. It is cognitive ability that counts. Pupils with similar CATs scores perform similarly regardless of social background or ethnicity. Pupils from more affluent backgrounds perform, on average, better than those from poorer homes because they are, on average,  brighter. Social background and even the quality of parenting make little difference by the age of 16. This is so contrary, not only to what politicians and the general public believe, and more importantly, to what they want to believe, that the facts do not see the light of day in mainstream educational discourse.

The evidence for this is in the latest book by international ‘intelligence expert’ James Flynnand is discussed in my review of his book.

Please now look again at the Cumbria schools scatter chart.

Note that the rank order of schools by GCSE exam results is completely different to the rank order of school effectiveness as determined from the chart. To find this out you have to do a little work. The regression line shows how, on average, GCSE results in Cumbria schools are driven by the mean cognitive ability of the Y11 cohorts in the schools. However some schools do better than average and those are to be found above the regression line. The greater the perpendicular distance above the regression line, the greater the effectiveness of the school.

You will see that on this basis ‘Gas Street Comprehensive’ (not its real name), with the worst GCSE performance, was more effective than the top performing school (the only selective grammar school in Cumbria).

So I wondered whether the same principle could be applied to the PISA international test results. To find out, I would need the equivalent of CATs scores. That was obviously a non starter so I thought about national mean IQ scores. In England we know there is a good correlation between parental education level and children’s performance at secondary school. The former is driven by parental cognitive ability/IQ. So if I could find data on national IQ scores I could use this as a proxy for the mean cognitive ability of the secondary age pupils in each country. Such national IQ data can be found here.

The principal author, Richard Lynn, is a respected international authority on the study of intelligence. Now some of my readers may begin to get uncomfortable. Discussion of ‘intelligence’ brings to mind the very nasty racist theories of the Third Reich. However unless the validity of ‘general intelligence’ is accepted it is very hard to get past discrimination-based explanations for the huge variety in cognitive competence that exists in the national population. This issue is discussed in detail in Part 1 of my book.

Like Piaget, Vygotsky and the developmental school of educationalists, I believe that intelligence is not fixed at birth, but is plastic. It can be significantly enhanced by the right sort of schooling  and subsequently by the autonomous decisions made by adults throughout life. This too is addressed in James Flynn’s new book. I discuss the importance of ‘Plastic Intelligence’ here.

The scatter chart at the end of this article is the result of my applying these ideas to the 2015 PISA international test results. I entered the PISA test scores for maths for every PISA participating country  on the Y axis, the national IQ data , expressed as percentiles on the X axis and used Excel to compute the regression line. The IQ percentile is the proportion of the national population with an IQ below that figure. For example the mean UK IQ is 100, which is the 50th percentile.

The resulting scatter chart  for the PISA countries is very like my chart for Cumbria schools. There is the same very strong correlation ‘R’ of 88 per cent between cognitive ability/IQ and exam performance.

It is hard to overstate the significance of this. Just as the Cumbria selective grammar school got the best GCSE results, the three Asian countries with the top IQs, Singapore, South Korea and Japan, came out top in PISA. However, all three are (like the UK and USA) below the regression line and therefore in the bottom half of the rank order of nationally effective education systems.

So which country has the most effective education system? The answer according to my chart is Ireland, followed by a number of European nations and Vietnam. The status of Ireland in my analysis results as much from the low national IQ score of 92 as from the good performance in the PISA tests.

So what are the guiding principles of the Irish education system? You can find these set out here in this 2007 government document. Eight later PISA has vindicated this Irish approach, which is very different to the marketised model of the English system and to the very didactic methods in the East Asian countries that top the PISA results tables.

The basic principles of the Irish education system are those of the developmental school (Piaget, Vygotsky, Shayer, Adey etc) as set out in Part 5 of . Learning Matters’. The parallels to the 2007 Irish government education approach will be obvious to every reader of my book. The following examples begin on p12 of the Irish publication.

The pedagogic principles of the Revised Curriculum which characterise the above learning processes are as follows: the child’s sense of wonder and natural curiosity is a primary motivating factor in learning; the child is an active agent in his or her learning; learning is developmental in nature: the child’s existing knowledge and experience form the base for learning; the child’s immediate environment provides the context for learning; learning should involve guided activity and discovery methods; language is central in the learning process; the child should perceive the aesthetic dimension in learning; social and emotional dimensions are important factors in learning; learning is most effective when it is integrated; skills that facilitate the transfer of learning should be fostered; higher-order thinking and problem-solving skills should be fostered; collaborative learning should feature in the learning process; the range of individual difference should be taken into account in the learning process; assessment is an integral part of teaching and learning.

It is clear that the more collaborative methods of teaching are the most effective.
Pupils also need to develop personal and group skills so that they may cope with the
social context for learning, and in order to retain knowledge most effectively.
Vygotsky’s “zone of proximal development” underpins many approaches to teaching
and learning in the primary school curriculum – those tasks too difficult for the child
to solve alone can be accomplished with the help of adults/peers, through instruction,
discussion and encouragement while the child internalises the ‘how to do’ bit of the
task as part of his/her inner speech for future reference. Hannan (1996), an independent
expert in how boys and girls learn, develops this idea further, and recommends a
“third/third/third” approach to proximal development, with pupils spending a third
of proximal learning time in friendship pairings/groupings, a third in single gender
non-friendship pairings and a third in mixed gender pairings, so that within one half
term everyone works with everyone else.

Ireland, whose children stand at the 30th IQ percentile outperform those of the UK at the 50th IQ percentile in maths. Ireland has no grammars schools, no Academies and no Free Schools. Surely the DfE should be looking for inspiration a short distance across the Irish sea rather than at the under performing  systems of East Asia.

The poor performance of these  PISA topping countries on my effectiveness chart follows from their very high national IQ scores.

With such intelligent populations their results would be much better still with the Irish/Estonian/Finnish/Canadian pedagogies rather than those in Singapore that got the Daily Mail so excited (shortly after PISA 2015 was published) about making the already miserable school experiences of English pupils even more unpleasant.

I have not commented on China. This is because the Lynn IQ data gives only a single figure of 100 for China, whereas the PISA results divide China into many regions with different levels of performance, which suggests large regional variations in IQ. Certainly Chinese pupils in the English education system produce mean CATs scores well above 100 (50th percentile). These regional differences in IQ may be relevant to the PISA performance of Wales and Scotland. If the mean national IQs in these countries are below 100 as they are in many of the poorer parts of England, then the anxiety about their PISA scores being below England may be misguided.

I give an explanation for the very high national IQs of China, Korea and Japan here.

My findings are so contrary to the mainstream view that they need to be checked. I would be happy to forward my full Excel file on receipt of an email. I will leave it to others to produce the scatter charts for Science and Reading.


Share on Twitter Share on Facebook

Be notified by email of each new post.


Roger Titcombe's picture
Tue, 13/12/2016 - 13:05

A very important contribution has today been posted by Child Development expert Dr Pam Jarvis. You can find it here

Here are some quotes from her article:

"The most pressing issue in education in many English-speaking nations is the lack of understanding of human development. The strategy of the Department for Education derives from a simplistic attitude that the earlier children enter education and the faster they are expected to learn, the better the outcome will be. The result is that learning is conceived as a “data dump” in which the teacher’s role is to act as a “memory stick” and children as human computers.

The development of cognition in human beings is not linear. Child development experts have known for approximately a century that there are distinct stages, and produced various models of this, for example Maria Montessori’s ‘planes of development’ and Jean Piaget’s stage model which outlines the increasingly abstract nature of human thought.


The PISA tests are constructed so that respondents cannot rely on the ability to regurgitate facts and figures; the respondent has to be able to use the knowledge that they have in a flexible fashion to apply it to a real-life example. Hence, a key reason why our children do poorly is that they have not been given enough time to develop the socio-cognitive skills required to apply knowledge in a flexible fashion. There are similar problems in other education systems in the UK which also make too early a start to formal teaching.

We can expect PISA results to get even worse in the future. The young people who took the PISA tests in 2015 went through their early years education before Michael Gove introduced his ‘accelerated curriculum’ – before teachers were pressured to drive children through a ‘too much, too soon’ teaching and learning process, inevitably based largely in highly pressurised rote learning. Not surprisingly, many are now quitting the teaching profession in order to protect their own mental health. This issue of both attracting and retaining teachers in England was raised in the 2016 PISA report.


So what of the education systems that did well on PISA 2016? The vast majority have school starting ages of six or seven. The two that start at five have a play-based early years curriculum, Aistear in Ireland and Ti Whariki in New Zealand. Finland, a reliable high performer for over a decade, has a school starting age of seven, a highly practical and applied manner of approaching teaching and learning, and no standardised testing system. England by comparison, has an average school starting age of four-and-a-half with some children starting school when they are barely four (although the age it supplies to PISA is five). It also applies standard assessment strategies to early years education from the very beginning.

In 2009, then Secretary of State for Education Ed Balls informed the nation’s parents ‘the sooner children start school, the sooner they close the gap on their peers’, while in 2014, the then Chief Inspector of OFSTED proposed that children should start school at two. Neither took any advice from those with expertise in human development before making such pronouncements to the national press."

In my local Labour Party branch we have a study group that will be meeting this evening to discuss Labour's, 'National Education Service'. This has been generally well received, but there are concerns about the section on 'Early Learning'. This is written primarily from the perspective of freeing a parent (usually the mother) to earn the money needed to pay the rent/mortgage (now increasingly the former) and feed the family. Astonishingly, free childcare is only available to a parent that already works 16 hours per week, and has already demonstrated the necessary  commitment  to prioritising the national GDP. This childcare is often low quality, provided by non expert, BTEC trained 'childcare' supervisors.

Quite apart from the many other serious reservations noted by Pam Jarvis, this clearly isn't good enough. No use looking to OfSTED for any useful contribution. Their 'Early Years' inspection regime is, as ever,  focussed on ensuring that DfE required systems are in place and that 'targets' are being met.

Roger Titcombe's picture
Wed, 14/12/2016 - 14:21

Since I published this article I have received a lot of email corresponsence from internationally respected academics. The general thrust of this has been to confirm the general validity of my approach. However Richard Lynn has informed me that my national IQ data are out of date and has provided me with the update published in the appendix of his book, ' Lynn &Vanhaned Intelligence: A Unifying construct for the S ocial Sciences,'

I have to emphasise that I am using these IQ data for the purpose of coming to sound conclusions about the interpretation of the 2015 international PISA test results. Others may use the data for other purposes and come to conclusions that I do not support. However, the IQ data now come fully referenced as to source and include updates resulting from the Flyn effect, along with confirmation that the Flynn increase in IQs over time has ceased for pupils in the UK, as noted in Section 5.10 of, 'Learning matters'. (I put this down to the degradation of the English education system caused by marletisation).

Most usefully the updated IQ data are given for various ages. In my corrections I will use the values, not above, and as close to age 15 as possible. This removes the weakness in my analysis of using adult IQs as a proxy for the average national IQs of the students taking the PISA tests.

I have also been advised that my scatter chart presentation my be too challenging for some readers, so I am now working on a different way of presenting the data. I will obtain the regression line formula from Excel and use this to find 'predicted' PISA maths scores by substituting the national IQ values into the formula. I will then subtract these predicted scores from the actual PISA scores to obtain a 'residual' for each country. Half of these residuals will be negative because the actual scores fall below the regression line. I am confident that this will still apply to the East Asian countries that head the 'raw scores' table and also the UK and the US.

I will then produce my own 'national education system effectiveness' table headed by the country with the highest residual. I am not sure this will still be Ireland, but that country will certainly be up there.

So, I must now get on with it. I would be really happy to share my Excel file with anybody that wishes to check my findings or has a better way of processing the data. Just email me.

Add new comment

Already a member? Click here to log in before you comment. Or register with us.