VoxEU Column Education

Curriculum and ideology

Schooling changes are associated with ideological ones but it is difficult to claim a causal relationship. This column attempts to analyse the causal effect of curriculum changes in China on shaping preferences of students. The new curriculum moves one’s belief about democracy by about 25% of a standard deviation in the direction desired by the government. The findings suggest the state can use education to promote socially-useful beliefs and cultivate good citizenship.

Education shapes young minds. Contemporary debates rage on whether it also shapes people’s political views, attitudes, and their values. Examples range from teaching of evolution in US schools, to the role of madrassas in the Islamic world, and the coverage of World War II in Japanese history textbooks. In 2012, an attempt to introduce a mainland Chinese curriculum into Hong Kong schools led to tens of thousands of people taking to the streets in protest.

Scholars across the social sciences have argued that schools play an important role in shaping political attitudes:

  • From the formation of a French national identity in the 19th century (Weber 1976);
  • To the cultivation of citizens capable of participating in the US democratic system in the 19th and 20th centuries (Dewey 1916, Lipset 1958);
  • To the production of masses who fit into the capitalist or socialist systems of the 20th century (for the former, see Bowles and Gintis 1976; for the latter, see Lott, Jr. 1991).

The idea that schooling can be used to mold – more pejoratively, ‘brainwash’ – children's views of their social, economic, and political environment is a powerful one, which resonates across time and space.

Yet, despite this resonance, and despite striking examples of schooling changes being associated with ideological changes, it is extremely difficult to determine whether schooling plays a causal role in shaping beliefs or if, instead, changes in curriculum simply coincide with other social, political, or economic changes which themselves shape preferences.

A reform in China

Starting in 2004, China introduced a nationwide reform of its curriculum, including significant changes in the textbooks used by students in senior high school (grades 10 through 12). This reform was shaped by a series of documents issued by the Ministry of Education and the State Council (the highest administrative body in the Chinese government), which reveal the importance attributed by Chinese officials to the textbooks and their ability to shape citizens’ views. One of the authors of the new textbooks wrote,

“Writing the Politics textbook is an act at the state level, rather than an academic activity of the individual author ... With a large readership, it will influence an entire generation of young people.”

The new textbooks were rolled out in a staggered manner across China’s 29 provinces between 2004 and 2010. In different years in different provinces, entering cohorts of high school students began studying from the new textbooks, preparing for a college-entrance exam (gaokao) that would be based on the new materials. Cohorts already in high school continued studying the old textbooks, and were tested on these. Map 1 shows the dates when the new curriculum was introduced in each province.

Map 1. Years of introduction of the new curriculum

In recent work (Cantoni et al. 2014), we attempt to identify the causal effects of this new curriculum. The staggered introduction allows us to compare the attitudes of adjacent cohorts of students who were exposed to different curricula within a particular province, while accounting for the cross-cohort differences that existed even when the curriculum was the same across those cohorts.

The objectives of the reform

We matched the goals of the curriculum reform articulated by the Ministry of Education and the State Council in official documents to changes in textbook content across curricula, and also to changes in the frameworks structuring the high-stakes gaokao exams. We identified the following objectives of the reforms:

  • Students should learn about Chinese democracy and political participation.
  • Students should learn about the importance of the rule of law for legitimising the Chinese government.
  • Students should study the “Three Represents” ideology expounded by Jiang Zemin, e.g., the importance of extending power to new segments of society.
  • Students should develop an appreciation for traditional Chinese ethnic heritage.
  • Students should understand and appreciate Chinese (non-market) economic institutions, e.g., the virtues of state-owned enterprises.
  • Students should be conscious of environmental issues.
Surveying Chinese students

To identify the impact of the new curriculum on the attitudes it was designed to shape, we conducted a novel survey among nearly 2,000 Peking University undergraduate students. The survey covers four cohorts of elite Chinese university students who entered high school between 2006 and 2010, drawn from 29 provinces. For each one of the textbook reform’s goals listed above, we designed a series of survey questions aimed at identifying changes in relevant attitudes.1

We find that the new curriculum was often successful in changing students' preferences and beliefs regarding important issues. Students who were exposed to the new curriculum were statistically significantly more likely to view China’s political system as democratic; to trust government officials; to favour extending political influence to groups outside the CCP (in line with the Three Represents); and, to view an unconstrained market economy with scepticism. We do not find that the new curriculum significantly changed Han Chinese students’ attitudes toward minorities (or minorities’ views of themselves); nor did the new curriculum cause students to favour policies protecting the environment, perhaps because these can be seen as opposed to economic growth – another high priority of the government.2

How big of an impact?

The magnitudes of the effects we observe are substantial. We estimate that around 20% of students who would not have held the government’s desired views in the absence of exposure to the new curriculum were persuaded by it. This makes schoolbooks a much more effective tool of persuasion than other means of so-called ‘oblique’ transmission of beliefs, such as advertising or mass media.3

We can also compare the effect of the new curriculum to the association between attitudes and students’ socioeconomic or demographic characteristics that might plausibly affect one’s views and attitudes. Our estimates imply that studying under the new curriculum moves one’s belief about democracy in China by about 25% of one standard deviation in the direction desired by the government. In comparison, having parents who are Communist Party members has an effect of only 10% of a standard deviation.

What lessons can we draw from this reform?

It is natural to wonder what our results imply about attempts by the state to shape attitudes in other cultural and social settings, e.g., in the US or in Western Europe. On the one hand, one might think that our findings represent the upper bound of what the state can achieve with a curriculum reform. After all, we study a project of an extremely effective Chinese state, and we study the hyper-competitive Chinese students who achieved nearly-perfect scores on their college entrance exams to attain admission at Peking University. On the other hand, Peking University students are known to be especially critical of their state, and willing to speak out against it, having been at the centre of pro-democracy movements, and continuing to openly discuss attitudes that deviate from the Party line. If the politically liberal students at Peking University, with access to microblogs and Weixin (WeChat) posts, had their views changed by their high school textbooks, we believe that a less critical population might be even more susceptible to the persuasive power of educational content.

Finally, it is worth considering the normative implications of our findings. While evidence of successful state brainwashing is frightening, our results also suggest that the state can use education as a force for the good, by promoting socially-useful beliefs. One hopes that, as Dewey and Lipset discussed decades ago, schooling will be used to cultivate good citizenship and sustain the working of a democratic state. Our findings suggest that it is within the power of education to achieve these aims.


Bowles, Samuel and Herbert Gintis (1976), Schooling in Capitalist America: Educational Reform and the Contradictions of Economic Life, New York, NY: Basic Books.

Cantoni, Davide, Yuyu Chen, David Yufan Yang, Noam Yuchtman, and Y. Jane Zhang (2014). “Curriculum and Ideology,” NBER Working Paper 20112.

DellaVigna, Stefano and Matthew A. Gentzkow (2010). “Persuasion: Empirical Evidence,” Annual Review of Economics, 2(1), September, 643-669.

Dewey, John (1916), Democracy and Education: An Introduction to the Philosophy of Education, New York, NY: Macmillan.

Lipset, Seymour Martin (1958), “Some Social Requisites of Democracy: Economic Development and Political Legitimacy,” The American Political Science Review, 1959, 53 (1), 69–105.

Lott, Jr, John R (1991), “Public Schooling, Indoctrination, and Totalitarianism,” Journal of Political Economy, December 1999, 107 (S6), S127–S157.

Weber, Eugen (1976), Peasants Into Frenchmen: The Modernization of Rural France, 1870–1914, Stanford, CA: Stanford University Press.

1 It is important to note that we specifically asked students questions in a manner that did not look like a series of examination questions. Nor do the patterns of responses look like responses to examination questions. There is a great deal of variation; indeed, the variation in responses to questions about trust in government officials is actually greater within our sample than in the broader AsiaBarometer sample of individuals. This also confirms the general view of Peking University students as relatively liberal and outspoken.

2 Note that these results are robust to controlling for province and cohort fixed effects, and do not result from different (socioeconomic) characteristics of the cohorts studying under the old and the new curriculum. In fact, along all observable dimensions the students in our sample who studied under the old and the new curriculum are virtually indistinguishable. This similarity in demographic characteristics across the two curricula suggests that we are able to identify a causal effect within our sample. For details on the econometric specifications and robustness checks, see the working paper.

3 For a theoretical definition of ‘persuasion rates’, and a review of the related literature, see DellaVigna and Gentzkow (2010).

5,459 Reads