A brief history of comparative education (and how the CES hopes to move us forward)

By Lucy Crehan

People have been attempting to learn from education systems ‘elsewhere’ for thousands of years. 

All the way back in Ancient Athens, Plato was admired the education policies of the city state of Sparta. He saw their education system as ingraining more discipline and order in its young men that the Athenian system did with its own citizens. 

Around the same time but on the other side of the world, Chinese thinking had a significant educational and philosophical influence on neighbouring countries’ such as Japan, Korea and Vietnam. In particular, their meritocratic system of civil service promotion based on exams was of interest to other countries. The imperial examination system reserved top spots in the civil service for scholars who performed the most highly in their examinations of Chinese classics, rather than selecting candidates based on their connections or wealth. 

Jumping inelegantly from ancient times to the 19th Century, nation states in Europe were developing national education systems for the first time, and scholars and officials travelled to other countries to learn what their neighbours were doing, to inform the design and improvement of their own systems. 

One such traveller from France, Marc Antoine Jullien, is an important figure in the history of comparative education, variously referred to as the ‘father of comparative education’ the ‘grandfather’ of comparative education, and also ‘the first scientifically minded comparative educator’. 

He spent years travelling across Europe studying education, and took a systematic approach to his studies. In his own words, his purpose was 

“To build up, for this science [comparative education], as has been done for other branches of knowledge, collections of facts and observations arranged in analytical tables so that they can be correlated and compared with a view to deducing therefrom firm principles and specific rules so that education may become virtually a positive science instead of being left to the whims of the narrow-minded, blinkered people in charge of it or diverted from the straight and narrow path by the prejudices of blind routine or by the spirit of system and innovation” (Gautherin, 1993, p6). 

Positivism, from whence the term ‘positive science’ comes, refers broadly to the position that knowledge, or facts, should be based on observation and experiment and the search for law-like regularities. Science should only be based on what is empirically experienced or verified through deduction. That is what Marc Antoine Jullien wanted for education. 

His approach to achieve this was that, again in his own words: 

“A series of questions on each branch of education and instruction, drawn up in advance and classified under uniform headings, would be given to intellectual and active men of sound judgement, of known moral conduct, who would search for solutions in public and private educational institutions, which they would have the mission of visiting and observing on different points.” (Philips & Schweisfurth, 2014, p33) 

Essentially, what he is suggesting there is that you would organise a team of people (of sound moral conduct), who would visit various countries and study their education systems based on categories and uniform headings that he gave them. There were a list of questions that he had about each country, and I’ll share just a few (from ibid. p34): 

“What is the proportion of the total number of these students to that of the population of the commune or the district?” In other words, what proportion of children attend school? 

“Approximately how many students are grouped under a single director or teacher?” In other words, what is the typical class size? 

“At which age are children admitted to primary schools?” In other words, what is the school starting age? 

These questions about access, class size, and school starting age are still those being asked and debated by modern scholars. 

It was in the first half of the 20th Century that comparative education became an academic field of study. It has been described as the period of prediction in the history of comparative education, because the study of other countries’ systems moved from simply descriptions and policy borrowing, to predictions about the likely success of borrowing, based on careful studies and comparisons of the social and cultural conditions in which education takes place. More thought and more care was given to the practice of transferring ideas or policies from one place to another. 

In the second half of the century, the call for a more scientific approach got even louder. There was a call for better methods, and a more systematic organization of the emerging field, but no universal agreement on what these methods should be, or how it should become more systematised. At the same time, the development of international organisations with an interest in education, such as UNICEF, UNESCO and the World Bank, created even more of an interest in comparative education, and a drive for it to be more scientific. 

At the end of the 1960s, Harold Noah and Max Eckstein wrote an influential text called – ‘Towards a Science of Comparative Education’, in which they set out an approach with the end goal of discovering causal relationships between educational outcomes and the educational and societal inputs which led to them. This approach suggested the analysis of educational data from many countries – and the more countries were included in the analysis, the more scientifically reliable the findings would be. 

Bear in mind, this was at a time when the emerging potential of computers was being realized, and the possibility of crunching large amounts of educational data contributed to the optimism about the possibility of comparative education becoming a positive science. 

Not everyone was optimistic though, and some disagreed with the whole project. Most notably, Edmund King, an eminent comparativist working around the same time as Noah and Eckstein, took a very different approach to his comparative education studies. He rejected the supposed neutrality and objectivity of the scientific method, believing instead that a subjective understanding of the viewpoints of students and teachers was both important to study, and relevant to policy. He put his fluent French, Italian, Spanish and his working knowledge of German to good use in examining the moral assumptions of education systems, and the different perspectives of those who work in them, along with their achievements and challenges. 

These differences of opinion about the purposes of comparative education are indicative of two very broad perspectives on what the objects of study in this field actually are. I’m going to draw on the work of Leon Tikly at the University of Bristol to characterize these different points of view, and the points of view of many scholars since, into two broad theoretical positions (Tikly, 2015). These two positions represent two different ways of seeing the world, and they take different stances on both ontology (what exists) and epistemology (how we should find out about our subject of study). 

What are these two broad positions then? 

The first is Empiricism. It is related to positivist science, which we’ve spoken about previously, in that it’s based on the assumption that science can ‘discover’ generalisable laws, replicable findings and reliable predictions, because there is an objective reality to discover, and events are related to each other in predictable ways. 

In the field of education, empiricism is the position which asks the question “what works?”. It values medicine-inspired research methods like randomised control trials, and econometric analysis, above other types of research. By contrast, it pays less attention and gives less weight to qualitative research. This is known as its epistemology – the ways in which empiricists think knowledge and facts should be discovered. 

The purpose of studies grounded in this empiricist position is often to identify causal links between an input and an output, or to create a model which reflects reality. For example, to discover the attributes of teachers which predict high performance in students. 

The second major ontological position, which incorporates a range of related positions such as social constructionism, ethnomethodology, and postmodernism, is known as Interpretivism. 

Interpretivism is a more relativist position. It emphasizes subjective meaning, and the role of language in constructing different versions of reality. What is real is subjective, and dependent on your context. This might sound strange if you’re referring to physical objects – a chair is a chair no matter what you call it – but makes more sense in the social sciences. What ‘knowing’ means for example, might depend on your culture. 

Researchers coming from this perspective therefore take quite a different epistemological approach, and focus on qualitative and participatory methodologies, like interview and ethnography, in which a researcher goes and lives or works with the people that they are seeking to understand. They don’t ask “what works?” or “what causes what?”, but instead ask questions like, “What is this community’s understanding of moral education?”. 

Our approach at the Centre for Education Systems is actually underpinned by a third perspective, which overcomes some of the problems inherent in empiricism and interpretivism. 

Let’s look at these problems first of all. We’ll start with empiricism and its preferred methodology, randomised control trials (RCTs). RCTs in education describe experiments in which students are randomly assigned to different conditions, for example, either a class size of 15 or a class size of 30. Everything else is kept the same as far as is possible, and then some measurement of progress is taken at the end of the trial to see whether the different conditions had different effects. If the average score of students in the smaller classes, however this is measured, is statistically significantly higher than the average score of the students assigned to the large class size condition, then the smaller class sizes are assumed to have caused the higher scores, and reducing class size is considered to be an effective intervention. You can see why this is a preferred approach of the empiricist position because – with a big enough sample size – this allows us to deduce a general law that a smaller class size leads to higher results, all else being equal. 

Now, one of the problems with this approach is that last little phrase - all else being equal. This approach to educational research seeks to look at just one variable at a time, and to see how that individual variable makes a difference to children’s learning. The trouble is, the effectiveness or otherwise of a particular variable such as class size may depend on other variables that have been kept the same – the teaching style for example. It might be that class size doesn’t make a difference if the teaching style stays the same, but a smaller class size combined with pedagogical changes that make the most of having a smaller class make all the difference, in a way that neither change by itself would. 

In any educational situation, the outcomes are jointly caused by many existing conditions – not just a single variable. Some may be more significant causal factors that others, but it will rarely be the case that only one variable is causing an outcome, and would continue to do so irrespective of the other conditions. This is difficult to model (though economists will try). To be educationally useful, you need to study the situation and see which different variables are working together, and in what way, to bring about a certain outcome. RCTs are useful clues as to what might be causing something, but in my opinion, they are not enough by themselves to be educationally useful. Nor are simple laws accurate descriptions of how the world works in education. 

This relates to a second problem with the empiricist, or positivist goal of coming up with generalizable laws in the social sciences. Even if an RCT strongly suggests that a particular intervention is leading to a positive outcome, and finds that using a particular reading programme leads to higher reading scores on average, this will hide variation within the sample. In other words, out of 300 children in the ‘treatment group’, i.e. those who participated in the reading programme, 280 may do significantly better that the matched control group, but 20 may do significantly worse. To say that the programme ‘works’ is missing something – and it is relevant and useful to conduct more qualitative enquiry to investigate why it didn’t work for these twenty individuals, and what can be done for them. 

Again, this doesn’t mean that RCTs aren’t useful, but rather that they don’t give rise to universal laws about ‘what works’ in education. The alternative view set out, of using them as a useful indicator and to direct further more qualitative investigations, is held by an ontological position known as critical realism, which is the approach the Centre for Education Systems takes. Put very simply, rather than asking “what works?”, a critical realist would ask “what works, for whom, and in what conditions?”. Critical realism still seeks to understand an objective reality in the social sciences, but recognizes that there are many causal and contextual factors at play in any situation, and tries to tease these apart through a combination of quantitative and qualitative methodologies. 

So why are we taking a critical realist perspective rather than an interpretivist one, given the above criticisms of the empiricist approach? 

Because while we respect the work of interpretivists in understanding and explaining the beliefs of individuals in different cultural settings, we also believe that there is a ‘reality’ beyond our own perceptions, values and interactions, and that what is effective at reaching various desirable educational outcomes is not entirely context dependent. 

Particularly because learning, the main focus of education, involves brains, which are remarkably similar around the world and are prone to learn in a similar way. I propose this means that some areas of comparative education (teacher power structures, for example) are more culturally dependent than others (such as the effect of automaticity on problem-solving). And while we can’t come up with educational laws that work everywhere, there is value in generating descriptions of how structures and practices work in different places, and attempting to draw tentative conclusions about the relationships between context, mechanisms and outcomes in more general terms, to help inform good policy making. That is the work we seek to do at the Centre for Education Systems.