Article Text

Download PDFPDF

Stress and psychosomatic symptoms in Chinese school children: cross-sectional survey
  1. Therese Hesketh1,
  2. Yan Zhen2,
  3. Li Lu2,
  4. Zhou Xu Dong2,
  5. Ye Xu Jun2,
  6. Zhu Wei Xing2
  1. 1UCL Centre for International Health and Development,London, UK
  2. 2Institute of Family and Social Medicine, Zhejiang University, Hangzhou, PR China
  1. Correspondence to Professor Therese M Hesketh, UCL Centre for International Health and Development, 30 Guilford Street, London WC1N 1EH, UK; t.hesketh{at}


Objective The Chinese educational system is highly competitive from the start of primary school with great emphasis on academic performance and intolerance of failure. This study aimed to explore the pressures on primary schoolchildren, and to determine the relationship between these pressures and psychosomatic symptoms: abdominal pain and headache.

Design Cross-sectional survey using self-completion questionnaires.

Setting/participants 9- to 12-year-olds in primary schools in urban and rural areas of Zhejiang Province, eastern China.

Outcome measures Proportion of children with defined school-related stressors and frequency of psychosomatic illness.

Results Completed questionnaires were obtained from 2191 children. All stressors were common in boys and girls and in urban and rural schools. Eighty-one per cent worry ‘a lot’ about exams, 63% are afraid of the punishment of teachers, 44% had been physically bullied at least sometimes, with boys more often victims of bullying, and 73% of children are physically punished by parents. Over one-third of children reported psychosomatic symptoms at least once per week, 37% headache and 36% abdominal pain. All individual stressors were highly significantly associated with psychosomatic symptoms. Children identified as highly stressed (in the highest quartile of the stress score) were four times as likely to have psychosomatic symptoms.

Conclusions The competitive and punitive educational environment leads to high levels of stress and psychosomatic symptoms in Chinese primary schoolchildren. Measures to reduce unnecessary stress on children in schools should be introduced urgently.

Statistics from


Over the past three decades, China has witnessed dramatic socio-economic change, which has had considerable impact on the Chinese experience of childhood. Sustained economic growth has greatly increased the purchasing power of the majority of families.1 The One Child Policy has resulted in most city-dwellers being only-children, while most of the rural population have only one sibling.2 The aspirations of many parents, who had limited educational opportunities themselves, are now invested in their only-children.3 This, together with previously unheard-of possibilities for upward mobility, has led to a highly competitive educational system.4 In addition, strong traditional ideas of childhood persist, based on the Confucian traditions of respect for parents and elders, filial piety, obedience and discipline.5

What is already known about this topic

  • China has a highly competitive educational system from the start of primary school.

What this study adds

  • Chinese primary school children experience high levels of stress in the school environment and are exposed to frequent physical abuse at home.

  • Stress is strongly associated with the classic psychosomatic symptoms of headache and abdominal pain in these children.

  • Measures need to be taken to remove unnecessary stress in the school environment.

There are increasing concerns about the effects of all these pressures on the health and well-being of children. A number of studies in Chinese adolescents have shown high rates of anxiety and depression, which are thought to be related to stress within the educational system.6,,8 But the pressure starts in primary school, where, from the outset, there are frequent exams, public ranking of students by exam results, large amounts of homework and intolerance of failure. Children are also often expected to conform to strict discipline for the first time, after being indulged by parents and grandparents in their early years.9

In children, stress and unhappiness are known to manifest as psychosomatic symptoms, with the most commonly documented symptoms being headache and abdominal pain.10,,12 Stress in the school setting has been shown to be significantly associated with psychosomatic symptoms in schoolage children in Norway, Sweden and Finland.12,,15 But this relationship has never been explored in a Chinese setting, where there are growing concerns about the stressful school environment.

This study had two aims: (1) to explore the pressures on primary schoolchildren in urban and rural China, and (2) to determine the relationship between these pressures and the psychosomatic symptoms of abdominal pain and headache.


The study was carried out in urban and rural locations in Zhejiang, a relatively wealthy eastern coastal province with a population of 48 million. The urban location for the study was Hangzhou, the provincial capital; the rural location was a poor county in Quzhou prefecture in western Zhejiang.

Questionnaires were developed for completion by primaryschool children, based partly on findings of previous research7 8 12 16 and on the results of informal focus groups conducted with children and parents for the purposes of informing the questionnaire content. Psychosomatic symptoms were assessed by the question: ‘In the past year have you had the following problems?’ The specific items were headache and abdominal pain. Response choices were ‘everyday, at least once per week, at least once per month, rarely or never.’ Responses were dichotomised for analysis, with those reporting that they experience these problems at least weekly defined as frequent.

Seven indicators of school stress were created: enjoyment of school, worry about exams, feelings of pressure to do well, difficulty completing homework, fear of punishment of teachers, being physically bullied and corporal punishment at home. Responses were: always or frequently, sometimes, rarely or never. Frequently and sometimes were grouped to give dichotomised variables for analysis. To identify whether the stress items were clustered in certain individuals, a score of school stress was developed with frequently scoring 2, sometimes 1 and rarely/never 0, producing a range of scores from 0 to a maximum of 14. The questionnaire was piloted in one urban and one rural school, and amended according to feedback.

Conduct of the study

At present, around one-third of the total population of Zhejiang is classified as urban, so we aimed for an urban and rural sample split which reflected this. Nine schools, three in urban Hangzhou and six in rural Quzhou, were randomly selected from the list of registered schools at the Provincial Education Bureau. Four of the schools initially approached declined to participate, so the next ones on the list were included. In each school, a random sample of children was selected from the school roll from Year 3 to Year 6 (age 9–12) for inclusion in the study. Lower year groups were not included mainly because of the need to ensure the level of literacy necessary to complete the questionnaire. Questionnaires were completed by the children in the classroom setting without teachers present. Research assistants were present to assist with understanding of questions where necessary.

Ethical considerations

Permission to conduct the study in the schools was obtained from the individual head teachers. Informed consent was obtained from parents (through a letter sent home) and from the child participants themselves. Anonymity and confidentiality were guaranteed. Children were told there was no compulsion to participate and that they did not need to complete any questions they found difficult to answer. They were also given the phone number of a counsellor willing to discuss, in confidence, any issues of concern raised in the questionnaire. In fact, none of the children availed themselves of this offer. Ethical approval was obtained from the Institute of Child Health in London, and local approval was obtained from the Zhejiang Provincial Education Committee.


The frequency of school stressor variables and psychosomatic health variables were initially analysed by individual school, gender and area of residence. We dichotomised all the stress variables before developing a logistic regression model to analyse the association between the measures of stress, and the psychosomatic symptoms, while adjusting for age, sex, residence and parental education. The education level of each parent was combined to develop a dichotomised parental education score (high/low) for analysis. We also separately analysed the effect of family size.


Characteristics of the study population

There were 2191 completed questionnaires, a response rate of 80%. Most of the 20% were excluded for inadequate completion. The non-responders were more likely to be younger than the responders (p=0.02), but there were no other consistent differentiating characteristics. We also excluded seven children for chronic serous conditions (diabetes, epilepsy, congenital heart disease and leukaemia). The sociodemographic characteristics of the sample are listed in table 1. The large difference in proportions of only children between urban (88%) children rural (44%) is explained by differences in implementation of the One Child Policy in urban and rural areas, which generally allows for two children in rural areas but restricts couples to one in urban areas. As would be expected, parental education levels were higher for the urban children. A very high proportion of children (93%) were living with both biological parents.

Table 1

Sociodemographic characteristics by residence n (%)

Prevalence of stressors and psychosomatic symptoms

The frequencies for the individual school stressors and the psychosomatic symptoms (abdominal pain and headache) by gender and place of residence are listed in table 2. All stressors are common in boys and girls, and in urban and rural schools. Nineteen per cent of the children rarely enjoy school, with boys less likely to enjoy school than girls. Eighty-one per cent of the total, 86% of rural boys and 89% of rural girls, worry ‘a lot’ about exams. Pressure to perform at school ‘all the time’ was reported by 78%, with urban girls reporting this the most at 83%. Over a quarter (26%) of the children find it difficult to complete their homework ‘always or very often.’ Nearly three-quarters are afraid of being punished by their teachers at least sometimes. Nearly half the children (44%) claim to have been physically bullied at least sometimes in school, with boys much more likely to be the victims of bullying. Finally, corporal punishment appears to be commonplace in the home: 71% of the children are at least sometimes physically punished by their parents, with small differences between boys and girls.

Table 2

Frequency distribution of stressors, headache and abdominal pain by gender and residence

Relationship between stressors and psychosomatic symptoms

The logistic regression analysis (table 3) showed that all stress variables were highly significantly associated with symptoms of headache and abdominal pain. Being bullied was the individual item which produced the highest odds ratios for headache (2.7, 95% CI 2.3 to 3.1) and for abdominal pain (2.1, 1.7 to 2.6). While physical punishment at home ‘often or sometimes’ produced an odds ratio of 1.6, further analysis of physical punishment for the 10% who are often punished gave odds ratios for abdominal pain of 4.3 (CI 3.1 to 5.4) and for headache of 3.9 (3.0 to 4.8). The overall school-related stress score showed a gradient across the three levels of stress, low, medium and high, with high levels of stress (score 11–14) being strongly associated with headache (OR 5.6, 4.2 to 7.0) and with abdominal pain (OR 4.9, 4.0 to 5.9). It is of note that being an only child was not significant ly associated wit h headache or abdominal pain. Eight per cent (n=175) of the total sample scored a full 14 points, with the strongest association with headache (5.9, 4.4 to 7.4) and abdominal pain (5.3, 4.2 to 6.5). It is of note that being an only child was not significantly associated with having headache or abdominal pain.

Table 3

Effect of stressors on the frequency of headache and abdominal pain controlling for gender, age, residence and parental education


Previous studies on school-related stress and its impact on health have been largely confined to Scandinavian countries, as noted above. While a few studies have touched on educational pressure and psychological morbidity in Chinese adolescents, this is the first study, to our knowledge, which explicitly addresses this in a primary school setting. This is despite growing awareness that the Chinese school environment is highly pressurised, even in the primary-school years.

Many of the children in this study were experiencing high levels of stress. It is striking that levels of stress are high across all four subgroups: girls and boys and in rural and urban schools. The levels were also remarkably consistent across the nine schools, so this does not relate to the environment of individual schools but suggests a widespread problem within the educational system. The majority of the children in the study experienced pressure to perform at school, worried about exams, found the volume of homework difficult to cope with and were afraid that teachers would punish them. In addition, bullying occurred frequently, and corporal punishment at home was commonplace. We also found high levels of the psychosomatic symptoms of headache and abdominal pain. There were strong, consistent associations between stress measures and psychosomatic symptoms. To put this into some context, the 10- to 13-year olds in a Swedish study of school stress reported that 21% of boys and 30% of girls experienced headache, and 17% of boys and 28% of girls experienced abdominal pain at least once per week.12

We believe that these high levels of reported stress are partly a reflection of the highly competitive, highly pressurised educational environment for children in China today. Given that these are primary-school children, it is of particular concern, since, in many societies, primary school is regarded as a time of nurturing. But other factors in Chinese society may contribute to these high levels of stress. First, the existence of large numbers of only children in urban areas has been blamed, because parents are thought to put greater pressure on onlychildren, when aspirations and expectations are focused on one child.17 Indeed, the limited psychological literature in this area in the early years after the onset of the Policy emphasised this concern.18 19 However, later larger studies have found few differences between only children and those with siblings across a range of psychological and health measures.7 8 20,,22 So, our own findings here that family size is not associated with either stress measures or psychosomatic illness is consistent with these later studies. This is evident at two levels in the study: at the level of the individual, stress and psychosomatic symptoms are not more common in the only-child than in sibling children, and at the population level, urban children, who are growing up in a predominantly one-child environment are no more prone to stress and psychosomatic symptoms than rural children.

Second, the pressures of urban living, and the rapid change which has been associated with it, have been implicated as a source of stress.23 In children, it is thought that academic aspiration is higher in urban areas, and hence stress levels are likely to be higher,17 but our findings that stress levels are higher in rural areas contradict this. We believe this may reflect the high value placed on education in Chinese society, urban and rural, and the widespread belief in the possibility for upward social mobility through education. Rural schools are generally disadvantaged in terms of their facilities and ability to attract the best teachers, thus placing additional pressure on rural children, who will ultimately have to compete with their urban counterparts for highly valued places in tertiary education.

Third, strict and punitive parenting may be a factor leading to high levels of stress. Only 27% of the children in the study were rarely or never physically punished, and those children who are very frequently punished are much more likely to manifest stress and psychosomatic symptoms. Hence, physical punishment at home is undoubtedly a contributing factor to high stress levels.

Finally, social relations between children may contribute. We found that nearly half of all children had experienced physical bullying, and as an individual stressor, bullying was found to have the strongest association with psychosomatic symptoms. The psychological effects of bullying have been widely studied, and a recent meta-analysis of studies found that there was a strong association between psychosomatic symptoms, especially headache and abdominal pain, and both being bullied and bullying.24 Bullying, of course, also occurs outside the school environment, and so its control is not just the responsibility of schools but a wider and more complex societal issue requiring complex interventions.

The study has a number of limitations. The cross-sectional nature of the study design means that the causal direction of the association cannot be assumed. In other words, it is possible that children with poor health, frequent headache and abdominal pain, may be less able to perform well at school, leading to increased perception of stress. We also did not ask about the home situation, except in relation to physical punishment, and home influences are likely to have an influence in terms of increasing or mitigating stress. The quantitative approach, with limited responses allowed for each question, is also somewhat crude and cannot elucidate the complex relationship between causal factors, stress and psychosomatic symptomatology. The next step would be a longitudinal study to confirm the direction of causation, and an in-depth interview study with a particular focus on those children with high stress scores, with exploration of the home situation as well. In addition, the study was limited to only nine schools in one province, with obvious issues of representativeness for the wider population. However, measures of stress and psychosomatic symptoms did not vary greatly between the schools; the major differences in the schools related to sociodemographic variables, such as gender differences, number of siblings and parental education, which were closely associated with urban/ rural residence, and we controlled for this in the regression analysis. Finally, the stressors we identified came largely from discussions at focus groups, and while we did pilot the questionnaire for understanding and relevance, we did not formally validate it.

The findings of this study have a number of implications. While high levels of stress are highly undesirable in the shortterm, there may also be long-term consequences. It is well documented in other countries that children identified with significant emotional and behavioural problems are likely to have psychological problems into adolescence and adulthood.25 26 Other studies have shown high levels of anxiety and depression in Chinese adolescents, which have been in part attributed to the highly competitive academic environment.7 22 There must be particular concerns for the long-term mental well-being of the 8% of children who scored maximum points on the stress scale. This identifies a very vulnerable subgroup, who appear to be coping badly in the school environment.

While competition per se cannot be eliminated in school, efforts can be made to reduce its more negative effects. Specific efforts to reduce school-related stress have been attempted elsewhere with mixed results.27 28 Much of the stress in Chinese schools is unnecessary and has simply become incorporated into the system. It is easy to address issues such as the frequency of exams, the common practice of public posting of ranked exam results and the sheer volume of homework. There are some schools already adopting some of these changes, including three of the schools in our study. More difficult is changing the long-held attitudes of parents and teachers. They need to be made aware that punishment may not improve performance and may indeed be highly detrimental in the short term and the longer term, and that encouragement should be the focus in this young age group. In addition, the clear association between stress and psychosomatic illness, which we identified, will highlight a hitherto poorly understood area for Chinese doctors. There is almost no awareness of the possibility of psychosomatic symptomatology in children, who are almost always treated entirely symptomatically for conditions such as headache and abdominal pain, without any consideration of underlying psychological causation. We hope that in disseminating these results, we may raise awareness of the frequency and importance of psychosomatic symptomatology in Chinese children.



  • Funding This study was funded though a grant from the Wellcome Trust.

  • Competing interests None.

  • Ethics approval Ethics approval was provided by the UCL Institute of Child Health, Zhejiang Education Committee.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Patient consent Obtained from the parents.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.