HEALTH PROMOTION INTERNATIONAL Vol. 19. No. 3 © Oxford University Press 2004. All rights reserved.
Socioeconomic predictors of smoking and smoking frequency in urban China: evidence of smoking as a social function
School of Urban and Public Affairs, University of Louisville, KY, USA
Address for correspondence: Zhenfeng Pan, 426 West Bloom Street, School of Urban and Public Affairs, University of Louisville, Louisville, KY 40208, USA, E-mail: terry3232{at}hotmail.com
SUMMARY
A number of previous studies have documented the worsening smoking problems in China. This paper identifies several key socioeconomic variables associated with smoking among urban working adults and calls researchers' attention to the important function of cigarette smoking as a social connection builder in China. This study argues that a smoking prevention and education program should place more emphasis on making smoking socially unnecessary and/or unacceptable because of the continuing increase in the number of smokers and the low desire to quit among smokers. The findings also suggest that while socioeconomic characteristics are excellent in predicting whether a person smokes or not, they are dismal in predicting smoking frequency among working adults.
Key words: adult smoking; smoking; smoking behavior; urban China
INTRODUCTION
In a study of cigarette smoking and sick leave in an industrial population in Shanghai (Wang and Dobson, 1992
), the authors claim that ‘cigarette smoking is one of the most preventable factors contributing to illness, disability and death in the world’. The limited research on smoking in China has painted a grim picture: Holden argues that ‘if current smoking patterns persist, tobacco will cause more than a third of all deaths among middle-aged Chinese men over the next few decades’ (Holden, 2001). Yang et al. predict that as the world's largest producer and consumer of tobacco, 2 of the 7 million deaths related to smoking will occur in China (Yang et al., 1999).
The widely cited research by Yang et al. reveals that smoking in China continues to be more prevalent among men than women, although the age of smoking initiation has declined by 
3 years for both men and women, from 28 to 25 years of age (Yang et al., 1999). They conclude that there is an urgent need for smoking prevention and cessation efforts in China, and tobacco control initiatives are needed to maintain or decrease the currently low smoking prevalence among women. Using descriptive statistics, Gong et al. studied the prevalence, characteristics and attitudes surrounding smoking in the Minhang district of Shanghai (Gong et al., 1995). Their research showed that 88.4% of respondents believed smoking was harmful for both the smoker and those exposed passively to the smoke, but only 14.1% of male smokers reported a desire to quit smoking. The researchers believed that the low quit rate and the low desire to quit were two of the most serious problems among smokers.
Sun and Shun are among the few academics to have studied smoking behavior across different socioeconomic groups in the Chinese workplace (Sun and Shun, 1995). Their analysis of differences in predisposing factors, enabling factors, reinforcing factors and smoking behavior among white- and blue-collar workers in Guangzhou suggested that significant differences exist between these two groups of workers. They reported that white-collar workers exhibit more positive predisposing factors towards smoking and consume more cigarettes than blue-collar workers.
On the other hand, Hesketh et al. and Li et al. found a very low prevalence of regular smoking among Chinese youths from 12 to 19 years of age (Li et al., 1999; Hesketh et al., 2001). The strongest factors associated with smoking among youths were maternal smoking, paternal smoking, smoking peers or friends, low paternal educational level and a poor self-reported academic record. One interesting result is that 75% of respondents said that they obtained cigarettes from home, and only 10% of youths obtained cigarettes through self-purchase. The attitudes of these youths were encouraging, with <10% of them thinking that young smokers have a tough or glamorous image, although significantly more of the current smokers thought that smoking improves mood and increases confidence or popularity. In a separate study of smoking among secondary school students in Henan, Zhang et al. reported that the likelihood of tobacco use was significantly higher among those with peers, teachers or mothers who smoked (Zhang et al., 2000). They also found that positive smoking-related attitudes among students had a significant association with their smoking status. Again, the authors argued for a comprehensive smoking prevention program in the early teen years so that it can be effective.
Xiang et al. examined cigarette smoking among medical college students in Wuhan, China (Xiang et al., 1999). None of the female respondents reported smoking on a regular basis. The major reasons for first smoking were stress, curiosity and loneliness. Multivariate analysis also revealed that having a family member who smoked was significantly associated with cigarette smoking for the students. Xiang et al. argued that their research suggested a need to initiate comprehensive tobacco-use prevention programs among medical students (Xiang et al., 1999).
From a global perspective, Peto et al. predicted that the sharpest increase in smoking would be in China and in early middle age (Peto et al., 1996). About 5% of women and 75% men in China are occasional or daily smokers, which prompted Richard et al. to claim that although ‘Chinese men number 10% of the adults in the world’, they ‘smoke about 30% of the world's cigarettes’.
Although a number of researchers have studied smoking behaviors in China, current research has several weaknesses: First, the research mostly concentrates on special segments of China's population in specific cities, possibly due to a lack of representative samples of the whole population. Secondly, some of the research is only descriptive in nature and does not analyze the relationship between smoking and other contributing factors. Thirdly, previous research tends to jump quickly to the conclusion that a comprehensive, effective prevention/education program is needed without providing detailed strategies. Finally, although the smoking behavior of youth has drawn some attention from scholars, the smoking behavior of adults has not been adequately studied in China. Even though previous research has identified gender and occupation as important factors associated with whether an adult smokes or not, the typical social and economic characteristics of adult smokers have not been adequately identified and explained. This study attempts to identify the typical socioeconomic characteristics associated with cigarette smoking and daily smoking frequency of the urban working population in China.
METHODS
Data
A nationally representative sample based on the Household Income Project in 1995 in urban China was used for this study. The principal investigators (Khan and Riskin, 2001, Riskin et al., 2001) claim that the data set is one of the best available to examine urban China nationwide. Only a few cases have missing values for some of the key variables and the missing values are replaced with the mean of each variable. It should be noted that the data set is only representative of the official urban residents in urban China (Riskin et al., 2001). The so-called ‘floating’ population, or migrants to the city from rural areas, were not included in the data collection.
Dependent variables
Two dependent variables were used for this study: the first was smoking (the respondent smokes regularly = 1, no = 0) and the second was daily smoking frequency (number of cigarettes smoked per day). Logistic regression was used to analyze the predictors of whether the respondents smoke or not, while ordinary least square regression was used to analyze the relationships between socioeconomic predictors and daily smoking frequency.
Independent variables
1. Continuous variables
Continuous variables included in the analyses were: age of the respondent, years of working experience, years spent in present job (measurement of work commitment), urban residence tenure (measuring the number of years since the respondent obtained official urban residence, or Hukou), Party membership tenure (a measure of seniority within the Party), number of persons in the household (household size), total family financial assets (the total cash value in Yuan or RMB of the financial assets owned by each respondent's family), total family disposable income in Yuan, and the average income of the respondent from 1990 to 1994 in Yuan. These income variables were constructed so that the effects of individual and family income could be assessed. This study also considered the estimated minimum living standard, which was measured by estimating the amount of money in Yuan that the household required to maintain a minimum standard of living. This variable was included in the model to adjust the regional differences in the cost of living, which may potentially affect smoking behavior.
2. Dummy variables
Dichotomous (dummy) variables included in this study were: female (yes = 1, no = 0); married (married = 1, no = 0); work unit owned by the state (yes = 1, no = 0); permanent employee (permanent employee = 1, no = 0) (permanent employees enjoyed a number of benefits and some prestige that others could not obtain in the previous planned economy); and educated youth (yes = 1, no = 0). With respect to this variable, a large number of young students was sent to country villages to help with agricultural operations during the Cultural Revolution from the mid-1960s to the mid-1970s; most of them came back to the city after the Cultural Revolution. These people experienced the dramatic event of the Cultural Revolution during the Chairman Mao era and this variable was included to test if these people were more likely to smoke than others. Other dummy variables included in this study were: public-financed health care (health care financed by the state or government = 1, other = 0) (recent health care reform has cut health care benefits for many employees, therefore smoking behavior may differ along this variable); and work indoors (yes = 1, no = 0) (a simple measure of work environment).
3. Scale variables
The first scale variable used in this study was occupational rank of primary employment. This scale was a rank of primary employment according to the Chinese Census Bureau definition. A score of 9 represented the head of a business or major government or business branch, and a score of 1 represented an unskilled worker. This variable was designed to examine the relationship between occupational rank and whether the respondent smoked or not. The second scale variable was an educational scale for the respondents. This variable ranged from 1–7, representing the different levels of education attained by the respondents (1 = illiterate/semi-illiterate, 7 = college or better education). Each scale increase represented 3 years increase in education. This approximate measure of respondent education was used to test if education impacted significantly upon the probability of smoking.
Analysis
As mentioned before, logistic regression was used to analyze the dichotomized dependent variable, while linear regression was used to analyze daily smoking frequency. Since logistic regression analysis revealed that Communist Party membership had a powerful positive influence on whether the respondent smoked or not, Communist Party members were selected from the sample and the predictors of smoking were also analyzed and presented in Table 2. The Discussion will focus on the Exp(B) coefficients because these coefficients are much easier to understand and give an intuitive understanding of not only the direction of impacts, but also the strength of the impacts of the independent variables. The impact of an independent variable is positive if the Exp(B) coefficient is >1 and negative if it is <1.
|
RESULTS
From Table 1 we can see that only 8% of smokers were female, 89.1% of smokers were married, 32.9% were Communist Party members, 67% had public-financed health care and 90% were permanent employees rather than contracted workers. In addition, 76% of smokers had government-assigned jobs, which carry a number of benefits and some social prestige.
|
In Table 2, logistic regression for all working respondents suggested that females were much less likely to smoke in urban China as indicated by the extremely small Exp(B) coefficient for the female variable. The older urban residents tended to be less likely to smoke, as indicated by the negative impact of the age variable.
The large Exp(B) coefficient of the marital status variable suggested that being married significantly increased the likelihood of smoking by the respondent, as did Communist Party membership. Those who were permanent employees, worked in state-owned enterprises, and had government-assigned jobs were more likely to be smokers than others. On the other hand, having more education, a higher minimum living standard and an indoor job tended to decrease the likelihood of smoking by respondents.
Among the Communist Party members, females and older respondents had a smaller likelihood of smoking than others. Interestingly, Communist Party membership tenure, higher employment rank and urban residence tenure decreased the probability of smoking among the Party members, suggesting that the older Party members with greater seniority, either for health reasons or others, had a better lifestyle in terms of smoking. However, working in state-owned enterprises and having longer work experience tended to increase the probability of smoking. Similar to the general working adults, a higher estimated living standard decreased the likelihood of smoking among Party members.
The last statistical model examined the socioeconomic factors associated with daily smoking frequency among smokers in urban China. The OLS regression results indicated that marriage increased the number of cigarettes smoked by the respondents. Greater education significantly reduced the number of cigarettes smoked. Interestingly, although total financial assets and individual income were not significant factors associated with whether the respondents smoked or not, higher total family financial assets reduced the amount of smoking, while average individual income increased the number of cigarettes smoked daily.
DISCUSSION
The logistic regression results confirm previous research that men are much more likely to smoke than women in China, but the findings also lead to several other interesting conclusions. First, smokers are much more likely to belong to some special segments of the population. The significant effects of the variables suggest that smokers are most likely to be male, married, Communist Party members, permanent employees, and workers in state-owned enterprises. Previous research (Pan, 2004) has suggested that these socioeconomic characteristics are strongly associated with various benefits such as house ownership, better housing location and housing quality from employer or government in the planned or transitional economy.
Secondly, the contrast between the high ‘% correct’ value of logistic regression and the dismally low adjusted r2 of the OLS regression model suggests that whether a person smokes or not is strongly associated with socioeconomic factors, while the amount smoked is probably determined by other factors, such as the length of smoking history, the amount of pressure or stress in life, the affordability of smoking, etc.
Thirdly, some researchers may be tempted to conclude from the findings that those with characteristics associated with various benefits and prestige in the planned economy and/or transition economy are more likely than others to smoke because they can afford to. Or they smoke because they have better access to cigarettes since packs of expensive cigarettes are frequently offered to them as gifts or bribery. These explanations are certainly valid to some extent. But the statistically non-significant effects of total family financial assets, total family disposable income and average individual income in the logistic regression models imply that whether a person smokes or not doesn't differ significantly along these ‘money’ variables. This study contends that smoking may be better attributed to the special social function it serves in Chinese culture. In China, smoking serves an important social function: the connection builder. When people meet each other, it is a custom for everyone to offer cigarettes to signal respect and hospitality. Among known friends or acquaintances, cigarette offering is equally prevalent and important, and serves the function of reinforcing friendships or relationships.
For a non-smoker, it is very easy to suffer the agony and even humiliation of not accepting someone's polite but pushy offer of a cigarette because Chinese culture accepts repeated offerings of cigarettes as a gesture of hospitality. It is unknown how many Chinese adults choose not to quit smoking because of the advantage smoking brings directly or indirectly in their daily interactions with bosses, friends and potential connections. This study argues that the people who are married, Communist Party members, permanent employees and state-owned enterprise employees tend to be more likely to smoke because they are more heavily influenced by the traditional culture of building and using connections than others in the transitional economy. Connections are much more important for these people because most of the benefits they receive, including housing benefits, are obtained through the influence of their connections within their work units. Bian explained in detail how connection (Guanxi) works in the job allocation process in the planned and transitional economy in China, and the importance of connection in daily life since, as Bian argues, the Chinese live in a web of social relations that often determine the attainment and allocation of various resources (Bian, 1994). A byproduct of the unique culture of using cigarettes as connection builders is a group of adults that this study defines as ‘social smokers’: those who only smoke around their friends or others when cigarettes are offered because they feel they have to use smoking as a gesture of maintaining their friendships so they don't hurt the ‘feelings’ and damage the delicate ‘relationships’ among their friends, even though they have decided to quit smoking on any other occasion due to emerging smoking-related health problems.
Smoking prevention is a monumental task facing the Chinese authorities unless they can make cigarette offerings both socially unnecessary and/or unacceptable. Researchers (Gong et al., 1995; Peto et al., 1996; Mackay, 1997; Tomlinson, 1997) have called for smoking prevention programs in China; however, efforts so far have neither made smoking a social taboo nor convinced smokers to quit. This study argues that future smoking prevention in China may use different strategies for the two groups: on the one hand, adult smoking prevention/reduction programs should place a heavy emphasis on educating the public about the potential harm of the cigarette offering ‘custom’ and on making it socially both unnecessary and unacceptable. Specific efforts should be made to maintain the very low prevalence of smoking among women. On the other hand, the smoking prevention/reduction programs for Chinese youth should place more emphasis on the bad image of smokers, since previous research has identified low rates of smoking among youth, alongside a low proportion of young people thinking that smokers have a tough and glamorous image (Li et al., 1999; Hesketh et al., 2000). By reducing both initiation at an early age and the necessity of smoking for adults, future tobacco use in China may be much lower than it is today.
Implication of this study
It is unknown how many Chinese adults started smoking because they could not turn down the repeated offer of cigarettes from friends or strangers they met. It is also unknown to what extent the low desire to quit smoking is the result of the need to remain a regular or ‘social’ smoker for social reasons. The significance of this study is that it brings our attention to the importance of culture and custom in strategy formulation to prevent smoking in China, where smoking serves a crucial social function in daily interactions. Because of the importance of culture in smoking prevention, future research should try to examine the underlying reasons for smoking initiation and the low desire to quit among adults so that coping strategies can be more effective. As Yang et al. point out, smoking-related illnesses and deaths will probably only occur in the coming decades in China (Yang et al., 1999). Unless effective prevention and reduction programs are implemented, the sheer size of China's smoking population will create grave consequences for the country's health care system.
ACKNOWLEDGEMENTS
The author would like to thank Ms Carolyn M. Skinner for her kind and generous help in editing the language of this paper. Both of the income study projects were jointly funded by The Ford Foundation, the Chinese Academy of Social Science Institute of Economics, The City University of New York, the Leverhulme Trust, Columbia University East Asian Institute and the University of California (Riverside).
REFERENCES
Bian, Y. (1994) Guanxi and the allocation of urban jobs in China. The China Quarterly, 140, 971–1000.
Gong, Y. L., Koplan, J. P., Feng, W., Chen, C. H., Zheng, P., Harris, J. R. (1995) Cigarette smoking in China: prevalence, characteristics, and attitudes in Minhang district. Journal of the American Medical Association, 274, 1232–1235.[Abstract]
Hesketh, T., Ding, Q. J. and Tomkins, A. (2001) Smoking among youths in China. American Journal of Public Health, 91, 1653–1655.
Holden, C. (2001) Tobacco epidemic in China's future. Science, 293, 1761–1762.
Khan, A. R. and Riskin, C. (2001) Inequality and Poverty in China in the Age of Globalization. Oxford University Press, New York.
Li, X., Fang, X. and Stanton, B. (1999) Cigarette smoking among schoolboys in Beijing, China. Journal of Adolescence, 22, 621–626.[CrossRef][ISI][Medline]
Mackay, J. (1997) Beyond the clouds—tobacco smoking in China. Journal of the American Medical Association, 278, 1531–1533.[CrossRef][ISI][Medline]
Pan, Z. (2004) Housing quality differentials in urban China: evidence from two national surveys. International Journal of Social Economics, in press.
Peto, R. et al. (1996) Tobacco—the growing epidemic in China. Journal of American Medical Association, 275, 1683–1685.[CrossRef][ISI][Medline]
Riskin, C., Zhao, R. and Li, S. (2000) Chinese Household Income Project, 1995 [computer file]. Inter-university Consortium for Political and Social Research (ICPSR) version. Producer: University of Massachusetts, Political Economy Research Institute, Amherst, MA; distributor: ICPSR, Ann Arbor, MI.
Riskin, C., Zhao, R. and Li, S. (2001) China's Retreat from Equality: Income Distribution and Economic Transition. M. E. Sharpe, Inc., New York.
Sun, W. and Shun, J. (1995) Smoking behavior among different socioeconomic groups in the work place in the People's Republic of China. Health Promotion International, 10, 261–266.
Tomlinson, R. (1997) China's smoking epidemic grows. British Medical Journal, 315, 502–505.[ISI][Medline]
Wang, W. Q. and Dobson, A. J. (1992) Cigarette Smoking and Sick Leave in an Industry Population in Shanghai, China. International Journal of Epidemiology, 21, 293–297.
Xiang, H., Wang, Z., Stallones, L., Yu, S., Gimbel H. W. and Yang, P. (1999) Cigarette smoking among medical college students in Wuhan, People's Republic of China. Preventive Medicine, 29, 210–216.[CrossRef][ISI][Medline]
Yang, G., Fan, L., Tan, J., Qi, G., Zhang, Y., Samet, J. M., Taylor, C. E., Becker, K., Xu, J. (1999) Smoking in China: findings of the 1996 National Prevalence Survey. Journal of the American Medical Association, 282, 1247–1254.
Zhang, L., Wang, W., Zhao, Q. and Vartiainen, E. (2000) Psychosocial predictors of smoking among secondary school students in Henan, China. Health Education Research, 15, 415–423.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  Z. Pan and D. Hu Hierarchical linear modelling of smoking prevalence and frequency in China between 1991 and 2004 Health Policy Plan., March 1, 2008; 23(2): 118 - 124. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. Mao, X. Li, B. Stanton, J. Wang, Y. Hong, H. Zhang, and X. Chen Psychosocial correlates of cigarette smoking among college students in China Health Educ. Res., February 16, 2008; (2008) cyn002v1. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||