Social influences on Cigarette Smoking Among Mainland Chinese and Chinese Americans: A Comparative Study

Abstract

The purpose of this cross-national study was to determine the leading social influences on smoking among Chinese Americans and mainland Chinese adults with the aim of improving prevention and intervention strategies to reduce smoking rates. A comparative cross-sectional design was used and a stratified-cluster sampling technique was employed in identifying the study sample. An 80-item questionnaire was administered in Chinese to 1222 participants, 812 in China and 410 in the U.S. Data were analyzed using descriptive and inferential statistical tests. Results contain comparative data on social influences of smoking among Chinese Americans and mainland Chinese, as well as factors influencing smoking behavior which include cultural beliefs, gender roles, and family relationship dynamics. Grandparents were influential and significantly correlated with current smokers in China. Findings indicate differential effects of social influences on smoking behavior. Similarities and differences provide a basis for improving and/or developing smoking intervention programs of mutual benefit to China and the U.S..

INTRODUCTION

Smokers get tobacco from social sources, such as family, friends, or strangers (Centers for Disease Control, n.d.; Forster, Chen, Blaine, Perry & Toomey, 2003; Shive, Ma, & Shive, 2001). Much of the research on social influences of cigarette smoking has been conducted on adolescents where the peer group or social network is the most important social influence. This influence is particularly strong among adolescents of ethnic and racial groups and is generally regarded as being more powerful than parental smoking (Hu, Flay, Hedeker, Siddiqhi, & Day, 1995; Landrine, Richardson, Klonoff, & Flay, 1994; Nguyen, Gildengorin, Gregorich, McPhee, & Kaplan, 2008; Otsuki, 2009), even though the influence of the latter cannot be underestimated (Morris, Vo, Bassin, Savaglio, & Wong, 1993; Newman & Ward, 1989; Nolte, Smith, & O’Rourke, 1983; Wen et al., 2005).

Few studies have explored the social influences of smoking, or the cultural values and social relationships that have a direct impact on Asian adult smoking behaviors. A study which focused on school-age Asian adolescent smokers by Landrine et al. (1994) found an association among school-related problems, truancy, and smoking as well as peer pressure, and that these were the best predictors of smoking behavior. A study by Hu et al. (1995) found that whereas Asian adolescents reported having fewer smoking friends and parents than white and black adolescents, the effects of parents’ and friends’ smoking behaviors were the strongest for the former group. Unger and Chen (1999) identified a high level of acculturation as a risk factor for initiation and perpetuation of smoking behavior. Results of a study by Ma et al. (2003) among Asians in the Eastern region of the U.S. found that, among males, smoking behavior was influenced to a large extent by the smoking status of fathers and other male members of the families and the number of smoker friends. The study also found that although peer pressure might influence smoking behavior among adolescents, this form of social influence was not an important factor for Asian smokers ages 22–44.

A study that compared smoking influences among mainland Chinese adolescents in Wuhan, China and their U.S. counterparts in California concluded that peer pressure and perceived access to cigarettes are important factors in smoking initiation. The study found, however, that Chinese Americans were more likely to have friends who smoked and to perceive easy access to cigarettes than their counterparts in China (Unger et al., 2002).

Much of the extant data on social influences that engender smoking behavior among Chinese adults in the U.S. and in mainland China is either inconclusive or limited in scope. For example, in China, smoking is an accepted social activity, especially among men (Yang et al., 2004). Cigarettes are routinely offered or exchanged at social gatherings where various age and gender groups are exposed to the behavior itself, as well as, the effects of secondhand smoke (Chen, Unger, & Johnson, 1999; Ma, Lan, Toubbeh, & Zhai, 2004). Generally, smoking among women in China is not a socially acceptable behavior because of the behavior’s perceived association with low morals, or simply because it is culturally unattractive. In China, smoking behavior tends to be gender-specific despite the population’s routine and extensive exposure to tobacco products (Jenkins et al., 1997; Ma, Chu, Jackson, & Tsou, 2003; Yang, Fisher, Li, & Danaher, 2006).

Chinese Americans comprise the largest subset of Asians in the U.S. (U.S. Census Bureau, 2012); many are recent immigrants who have been either exposed to smoking or are smokers themselves. China has over 300 million smokers who are at risk of developing smoking-related illnesses, and who place others at risk of developing such illnesses. Both the U.S. and China are experiencing the adverse health and economic consequences of tobacco use and each has unique responsibilities in dealing with these effects and consequences. Comparative studies are essential not only to understand social influences of smoking behavior in situ, but also to benefit from mutual experiences with the ultimate goal of eliminating smoking and reducing its burden on the two societies. The purpose of this study was to examine similarities and differences in social influences of smoking among mainland Chinese and Chinese Americans with the aim of contributing to the knowledge base about these influences, to identify culturally appropriate factors that could be used in prevention and intervention strategies that may decrease overall smoking rates in China and the U.S.

METHOD

Procedures

A cross-sectional study was conducted to compare smoking related knowledge and behaviors among Chinese Americans and mainland Chinese. The total sample consisted of 1222 participants, age 18 and older, and was comprised of 812 Chinese residing in Nanjing and Chengdu, China, and 410 Chinese Americans residing in the Delaware Valley region of Pennsylvania and New Jersey of the U.S.

A stratified sampling design was utilized in selecting the representative Nanjing and Chengdu sample. Six community neighborhoods were identified in Nanjing and Chengdu. Within the identified neighborhoods, 868 households were randomly selected. One member from each household was then selected based on gender and age population distributions in each of the two cities. Of the 868 household members who agreed to complete the study questionnaire, 812 complied, representing a 93.4% return rate. A stratified-cluster proportional sampling technique was used for selecting the Chinese American sample (Ma, Shive, Tan, & Toubbeh, 2002). A current listing of Asian American community organizations (N = 52) was identified by the Center for Asian Health at Temple University. These organizations were located in geographic areas of Pennsylvania and New Jersey which covered Chinese across ages and socio-economic status. Twenty-six organizations were randomly selected as clusters from the list of 52. One of the eligibility criteria is to limit organization participants to only one adult per household. A sample of 410 Chinese American adults age 18 and above completed the study, representing a participation rate of 96.3%. Efforts were made to ensure the Chinese American and Mainland Chinese samples were comparable. The investigation stratified both samples by age, gender and education level.

To ensure the completion of questionnaires and accuracy of information in China and the U.S., surveys were administered in a face-to-face format by personnel trained in procedures and protocols of the study. In China, the School of Public Health at Southeast University and the Huaxi School of Public Health at Sichuan University provided graduate and undergraduate student volunteers to assist; in the U.S., bilingual staff from the Center for Asian Health and Asian Community Cancer Coalition provided on-site translation, monitoring, and face-to-face facilitation.

Instrument

An 80-item questionnaire was administered to the U.S. and China samples. The questionnaire contained items that measured smoking knowledge, attitudes, social influences, perceived risks with regard to demographics, smoking status, smoking cessation, and advice from health professionals.

Pilot tests were conducted in the U.S. and China to determine the feasibility of the study, face and content validity, and internal consistency of the measures. In the U.S., the test was conducted among 50 Chinese American participants and in China (cities of Nanjing and Chengdu), among 35 participants. All participants in the pilot studies were excluded from the larger study. Improvements in the survey were made on the basis of participant feedback (Ma, Shive, Tan, & Toubbeh, 2002). To ensure the completion of questionnaires and the accuracy of the information, face-to-face interviews with participants were conducted by personnel who had received prior training in procedures and protocols of the study. The amount of time required to complete the questionnaire was 20–25 minutes.

Data Analysis

Categorical and continuous demographic variables were compared across China/U.S. datasets using chi-square two-sample t-test statistics for categorical and continuous measures, respectively. Within each dataset, previous and current smoking status (yes vs. no) for relatives and friends were compared by the participant’s current smoking status using the chi-square statistic. The number and percentage of friends who smoke were also compared by participant smoking status within each dataset using the two-sample t-test. Within the dataset, logistic regression models were used to assess predictors of current smoking status as a function of the following independent variables: gender, categorized educational level, marital status, dichotomized indicator of children under 18 living in the home, categorized age, number of friends who smoke, and current smoking status of father, grandmother, grandfather, and brother.

RESULTS

Demographic Data

Demographic characteristics are presented in Table 1. The sample was comprised in the following way: China - male, 58.3%; female, 41.7% female; U.S. - male, 53.8%; female, 46.2%).

Table 1.

Demographics

	China (n=812)	U.S. (n=410)	χ2	Total (n=1222)
Gender, %
Male	58.3	53.8		56.8
Female	41.7	46.2	2.2	43.2
Level of Education, %
Illiteracy	2.5	0.0		1.7
Under junior middle school	28.2	10.5		22.3
Senior high school	31.1	32.3		31.5
Vocational school	6.7	3.5		5.6
College or higher	31.5	53.7	336.6**	38.9
Marital Status, %
Married	74.2	69.1		72.5
Separated	0.4	2.0		0.9
Divorced	1.5	1.5		1.5
Widowed	2.0	3.0		2.3
Single	21.7	22.5		22.0
Others	0.2	2.0	19.5**	0.8
Children<18y at home, %	53.7	55.6	1.2	54.3
Age Group
13 – 17	4.9	7.5		5.8
18 – 22	7.1	6.0		6.7
23 – 40	47.9	37.7		44.5
41 – 60	31.7	35.7		33.0
>=61	8.4	13.2	17.3**	10.0
Age in years, (mean, SD)	39.3 (14.5)	41.4 (15.8)	2.3*	40.0 (15.0)

^*p < .05;

^**p < .01

There were similarities in the information reported by respondents in both countries. For example, overall information on marital status, level of education and age was comparable. The majority of respondents reported that they were currently married (U.S., 69.1%; China, 74.2%), and had a senior high school level of education (U.S., 32.3%; China, 31.1%). The mean age of U.S. respondents was 41.4 years old and China, 39.3 years. About 79% of the China sample was between 23 and 60 years of age and approximately 73% of the U.S. sample was between 23 and 60 years of age.

Current Smoking Status and Family Members

Table 2 shows the data by respondent smoking status for ever smoking and currently smoking family members and friends. There were significant differences in ‘family and friend smoking’ for non-smoking respondents and respondents who are current smokers. There were also differences in ‘friends who smoke’ between nonsmokers and current smokers within the China and U.S. samples and between countries.

Table 2.

Smoking Status by Country and Family Members

	China			U.S.
	Non-smokers n=367	Current smokers n=302	χ2	Non-smokers n=311	Current smokers n=99	χ2
Ever smoking family members, %
Father	52.4	64.1	10.0**	48.1	58.6	3.3
Mother	3.4	7.5	6.2*	12.0††	13.1	0.1
Uncle	21.6	28.4	4.5*	22.1	24.2	0.2
Brother	22.6	39.7	25.1**	25.3	30.3	0.9
Sister	0.5	1.6	2.2	1.9	6.1†	4.4*
Husband	33.0	4.7	88.2**	12.3††	4.0	5.6*
Wife	2.2	1.9	0.1	0.6	4.0	5.9*
Currently smoking family members, %
Father	37.1	33.8	0.8	12.5††	24.2	7.9**
Uncle	17.4	17.2	0.006	15.8	25.3	4.6*
Brother	16.1	32.8	25.6**	1.3††	3.0††	1.4
Sister	0.5	0.3	0.2	1.6	6.1††	5.7*
Grandpa	12.5	2.3	23.7**	0.3††	2.0	3.0
Husband	22.3	4.0	46.3**	3.2††	4.0	0.2
Wife	3.5	7.3	4.7*	3.2	1.0†	1.4
Friend smoking, %	61.6	85.5%	82.5**	48.3††	64.5††	7.4*
Friends who smoke (Mean, SD)	25.7 (27.8)	63.7 (27.3)	18.8**	4.0 (3.1)††	5.5 (2.8)††	3.7**

^*p < .05;

^**p < .01

^†p < .05 (Nonsmokers of China vs. Nonsmokers of U.S.) or (Current smokers of China vs. Current smokers of U.S.)

^††p < .01 (Nonsmokers of China vs. Nonsmokers of U.S.) or (Current smokers of China vs. Current smokers of U.S.)

Current smokers in China were more likely to report ever-smoking among fathers, mothers, uncles, brothers, and sisters than were non-smoking respondents. There was a significant difference between nonsmokers and current smokers with respect to ever smoking fathers, mothers, uncles and brothers. However, nonsmokers in China were more likely to report husbands ever-smoking than were current smokers (33% nonsmokers, 4.7% current smokers; p < .01).

There were significant differences between nonsmokers and current smokers of ‘currently smoking family members and friends’. Current smokers (32.8%) in China were more likely to report brothers who ‘currently smoke’ than were nonsmokers (16.1%), a ‘wife who currently smokes’ (7.3% current smokers vs. 3.5% nonsmokers; p < .05), and ‘friend who currently smokes’ (85.5% current smokers, 61.6% nonsmokers; p < .01). However, nonsmokers in China reported significantly more ‘grandfathers who currently smoke’ than did current smokers in the sample (12.5% nonsmokers, 2.3% current smokers; p < .01), and nonsmokers also were likely to report’ currently smoking husbands’ than did the current smokers in the China sample (22.3% nonsmokers, 4% current smokers; p < .01).

Similar to current smokers in China, current smoking Chinese Americans were also more likely to report ever smoking by fathers, mothers, uncles, brothers, and sisters than were nonsmokers in the U.S. sample; however, unlike the China sample, the difference between these in the U.S. sample was not statistically significant.

Ever Smoking Status and Family Members

There were significant differences in ‘ever-smoking’ by sister, husband, and wife in the U.S. Current smokers in the U.S. sample reported significantly higher percentages of ‘ever-smoking’ by sister and wife than nonsmokers (‘ever-smoking’ by sister: 6.1% current smokers, 1.9% nonsmokers, p <. 05; ‘ever-smoking’ by wife: 4% current smokers, 0.6% nonsmokers, p < .05). Nonsmokers in the U.S. reported significantly more ‘ever-smoking’ by husbands than did current smokers (‘ever-smoking’ by husband: 12.3% nonsmokers, 4% current smokers, p < .05). Among current smokers in the U.S., 24.2% reported a father ‘currently smoking’, while 12.5% of the nonsmokers reported a father ‘currently smoking’ (p < .01); 25.3% of current smokers reported an uncle ‘currently smoking’ and 15.8% of nonsmokers reported an uncle who was a ‘current smoker’ (p < .05); 6.1% of current smokers reported that a sister was a ‘current smoker’, and 1.6% of nonsmokers reported a sister who ‘currently smoked’ (p < .05). Current smokers in the U.S. also reported significantly more ‘friends who are current smokers’ than did nonsmokers (64.5% current smokers, 48.3% nonsmokers, p<.05).

Comparing the U.S. and China data, similar trends can be seen among those reporting ‘ever-smoking’ family members. Current smokers in China and the U.S. were more likely to report ‘ever-smoking’ for father, mother, uncle, brother, and sister than did nonsmokers. Nonsmokers were more likely to report ‘ever-smoking’ husbands. The only dependent variable that was significant for both the U.S. and the China samples was husbands who ‘ever smoke’ (China sample, ‘ever smoking’ husband: 33% nonsmokers, 4.7% ‘current smokers, p < .01; U.S. sample, ‘ever smoking’ husband: 12.3% nonsmokers, 4% current smokers, p < .05). With respect to the data on ‘currently smoking’ family members, the only dependent variable for which a similar trend was seen between the U.S. and China was for having a currently smoking brother, yet this variable was significant in the China sample only (China, ‘currently smoking’ brother: 16.1% nonsmokers, 32.8% current smokers, p < .01; U.S., ‘currently smoking’ brother: 1.3% nonsmokers, 3% current smokers).

Comparison of Friends Who Smoke

The difference in ‘friends who smoke’ between nonsmokers and current smokers followed the same trend and was statistically significant in both countries’ samples (China: 61.6% nonsmokers, 85.5% current smokers, p < .01; U.S.: 48.3% nonsmokers, 64.5% current smokers, p < .05). The mean for ’friends who smoke’ highlights the similarities and differences between the China and U.S. data. In both countries’ samples, current smokers reported significantly higher numbers of ‘friends who smoke’ than did nonsmokers, and the mean reported in China was much greater than in the U.S. (U.S. current smoker: 5.5 ‘smoking in friends’; nonsmoker: 4.0 ‘smoking in friends’, p < .01 and China current smoker: 63.7 ‘smoking in friends’; nonsmoker 25.7, p < .01).

There were statistically significant differences between the two countries for ‘ever smoking’ family members, ‘current smoking’ family members, proportion of ‘friends who smoke’, and number of ‘smoking friends’. Among ‘ever smoking’ family members, nonsmoking Chinese American mothers were significantly more likely to smoke than in China, but more husbands were likely to smoke in the China sample. Among ever smoking family members, current smoking Chinese American sisters were more likely to smoke than those in China. Among current smoking family members, nonsmoking Chinese American fathers, brothers, grandfathers, and husbands were significantly less likely to smoke than in China. Among ever smoking family members, current smoking Chinese American brothers and wives were significantly less likely to smoke than those in China. Significantly more friends smoke in China than in the U.S. among nonsmokers and current smokers.

Social Influence Correlates for Current Smoking Status

Table 3 shows the results of two logistic regression models: a China and a U.S. model. In the China model, ‘grandfather smoking (OR = 64.7, p < .05) and ‘grandmother smoking’ (OR = 2.03, p < .05) show a significant positive relationship with smoking. The number of ‘friends who smoke’ also has a positive relationship with current smoking; the more friends who smoke, the more likely the respondent was to be a current smoker (‘number of friends smoking’: OR = 1.03, p < .01). Age was also a significant factor associated with smoking: as age increased, the likelihood of smoking also increased. A positive association was also seen between gender and smoking, with men more likely to be smokers than women in our sample.

Table 3.

Final Logistic Regression Models of Social Influences for Current Smoking Status China Model U.S. Model

	China Model		U.S. Model
Factors2	OR	P	OR	P
Grandpa smoke	64.70	.013**	1.15	.874
Grandma smoke	2.03	.012**	3.15	.229
Father smoke	1.08	.788	2.20	.082
Brother smoke	1.21	.581	1.14	.744
Friend smoke	1.77	.050*	1.10	.770
No of friend smoke	1.03	.000***	1.23	.001**
Age	1.71	.003***	1.10	.548
Gender(female)	0.06	.000***	0.28	.000***
Constant	0.37	.303	0.38	.284

Note: 1. Independent variable: Current smoking.

²Dependent variables in the Logistic model

^*p < .1;

^**p < .05;

^***p < .01

In the U.S. model, only two factors were significant in influencing smoking. Number of ‘friends who smoke’ was positively associated with smoking. Respondents were 1.23 times more likely to smoke if they reported having ‘friends who smoke’ (p < .05). Gender was also a significant factor associated with smoking in this model. Being female was a protective factor for smoking (OR = 0.28, p < .01). Comparing the data for the China and U.S. sample, gender and number of ‘friends who smoke’ were the only two factors that were significant in both samples.

Discussion

This study found that social influences among current smokers and nonsmokers differ in both Chinese Americans in the U.S. and mainland Chinese in China. Whereas in the U.S. significant differences were reported for ‘ever smoking members’ for sisters, wives, and husbands, in China, current smokers were significantly more likely to have parents, uncles, brothers, and husbands who smoked than nonsmokers for ‘ever smoking members’ of the family. The former findings support previous studies by Ma et al. (2003) and Fang, Li, Stanton, & Dong et al. (2003). Results also suggest that the social influences of cigarette smoking are significant among Chinese Americans and mainland Chinese. However, the social influences of mainland Chinese are different from those of Chinese Americans; they tend to be much stronger. A possible explanation is the lack of health information in China about the hazards of smoking. Although the country has made several attempts to change public perception of smoking, its policies and their implementation lag behind those of the U.S.

The study corroborated earlier studies by Fang et al. (2003) regarding the relationships among age, gender and smoking. In China, cigarette smoking is considered both an accepted social activity and a part of social interaction, especially among adult males. These norms however, do not apply to women because of the association between smoking and low morals (Jenkins et al., 1997; Lai, Ho, & Lam, 2004; Ma, Shive, Toubbeh, Tan, & Zhao, 2003; Mao et al., 2009). This may explain, in part, why there was a greater proportion of ‘friends who smoked’ for nonsmokers and current smokers in the China sample than in the U.S. The significant difference in the relationship between age and smoking between the countries’ samples may be a function of China’s ban on tobacco in elementary and high-schools under Article 5 of the tobacco law of 1991 (Mackay, 1992).

The correlation between ‘having friends who smoke’ and smoking is similar in both countries and corroborates previous studies (Altman, Levine, Coeytaux, Slade, & Jaffe, 1996; Hu, Flay, Hedeker, Siddiqhi, & Day, 1995; Newman & Ward, 1989; Nolte, Smith, & O’Rourke, 1983; Pierce, Choi, Gilpin, Farkas, & Berry, 1998). In previous studies, as well as in this study, smokers were more likely to have friends who smoke, while nonsmokers were more likely to have non-smoking friends. The influence of social networks on smoking in China and the U.S. differ. Whereas previous studies have suggested that a high level of acculturation among Chinese Americans is associated with increased risk of smoking (Unger et al., 2002), our findings indicated an opposite trend. We found that grandchildren of Chinese American smokers in our sample were generally nonsmokers. This finding was not reflected in the China sample where the older the person, the more likely they were to smoke. Current smokers in the sample had grandfathers and grandmothers who smoked. These differences in social influences may be attributed to the differing roles of grandparents in the two countries. In the U.S., Chinese grandparents provide less daycare for their grandchildren on a regular basis. In China, however, grandparents, who are usually retired, are nearly always daycare providers for their grandchildren to help their own employed children. In so doing, they become models for their grandchildren.

The study revealed that the smoking behaviors of mainland Chinese and Chinese Americans are influenced, to a large extent, by their respective social networks. Cultural mores in mainland China are far more accepting of male smokers than of female smokers, hence the predominance of smoking among males. This trend appears to persist among Chinese Americans despite a change in social networks. There is, however, a difference in the rate of smokers among the latter that could be attributed to new and different social influences as environmental restrictions on smoking in the U.S. increase and enlightenment about tobacco use risks grow in the Chinese community.

China’s dual cultural standards of condoning smoking among males while associating loose conduct with smoking among females is curious, especially in the light of the fact that smoking among elder females is condoned. Filial piety may be a factor that could be addressed in prevention and cessation programs.

The role of Chinese elders (e.g., grandparents) in China and in the U.S. appears to have differential effects on smoking behavior, another important factor which may be considered in prevention programs. China’s economic upsurge has underscored the role of grandparents in nurturing the young while parents are at work, exposing the young to secondhand smoke and facilitating their use of tobacco. What lessons can be learned from these observations may be a subject for further exploration, particularly in China.

One limitation of the study is the numerous factors that are different between two cultures that can influence a given behavior, especially a health behavior. Further, methods which ask participants to determine the importance of a set of values by ranking for comparison cross-culturally may not always yield the most accurate and complete understanding of true value differences. For example, people in different cultures have divergent attitudes because they do not view a reality in the same way (Peng, Nisbett, & Wong, 1997).

Cross-cultural studies are studies of members of various groups that have had different experiences that lead to predictable and significant differences in behavior. We focused on two groups, ethnically and racially related, who share common cultural values but who reside on different continents. Chinese Americans represent the largest subset of Asian Americans, many of whom are recent immigrants who continue to hold onto traditional values and mores, despite the challenges associated with acculturation. Mainland Chinese face similar challenges within their own milieu—a societal transformation that is affecting the lives of over a billion citizens. Both our samples share a common burden - tobacco and the serious heath risks associated with it. Both countries can gain from mutual cooperation in addressing the current range of issues associated with tobacco production, distribution, and use. This study focused on the significance of social networks in, and social influences on tobacco use. Insights gained can contribute significantly to early prevention as well as intervention strategies.

Although it has been known that men tend to smoke more than women cross-culturally and there is a higher prevalence in China than the U.S., this study has examined more closely the social influences of smoking comparing U.S. Chinese with Chinese in China. Public Health professionals who are attempting to reduce the harm of cigarette use and improve on prevention and cessation can use this information to target social influences within the Chinese population more respectively, given the idiosyncrasies of each respective nation.

In conclusion, smoking behaviors of Chinese populations, both in U.S. and in China, are greatly influenced by their social networks, such as family members and close friends. Our findings suggest that any successful smoking prevention and intervention strategy in the future for Chinese populations must emphasize the substantial impact of social networks.