Journal of Youth and Adolescence

, Volume 46, Issue 6, pp 1262–1274 | Cite as

Tobacco Use by Middle and High School Chinese Adolescents and their Friends

  • Ling Li
  • Ting Lu
  • Li Niu
  • Yi Feng
  • Shenghua Jin
  • Doran C. French
Empirical Research
  • 226 Downloads

Abstract

Understanding the similarity of the tobacco use of youth and their friends and unraveling the extent to which this similarity results from selection or socialization is central to peer influence models of tobacco use. The similarity between the tobacco use of Chinese adolescents and their friends were explored in middle (880, 13.3 years, 399 girls) and high school (849, 16.6 years, 454 girls) cohorts assessed yearly at three times. Boys were more similar to their friends in tobacco use than were girls. Growth curve models revealed escalation of use during middle school and stable use during high school for boys, whereas models for girls could not be computed. Evidence of selection effects emerged from cross-lagged panel analyses revealing pathways from boys’ tobacco use to subsequent changes in their friends’ use; assessment of selection and influence processes could not be assessed for girls. The results from this study suggest that peer influence processes may differ for Chinese boys and girls and that further quantitative and qualitative research is necessary to understand these processes.

Keywords

Tobacco Friendship: China Peer influence 

Introduction

The tobacco use of adolescents in the United States follows a variety of trajectories (Nelson et al. 2015). In a longitudinal study of Minnesota youth (Bernat et al. 2008), most adolescents abstained from use (54 %), others experimented with tobacco (17 %) or smoked occasionally (17 %), with the usage increasing slightly during adolescence. Some regular smokers initiated use at a young age (7 %) (12–14 years), and increased their use such that they become regular users by age 16, whereas others did not start smoking until about age 14 (8 %) after which they become regular users. Most researchers have found similar trajectory subgroups (Nelson et al. 2015). Despite recent declines (Johnston et al. 2016), US adolescent tobacco use remains a serious health problem with approximately 3.6 % of eight grade and 6.3 % of tenth grade students in 2015 reporting use within the prior month.

Multiple theoretical models of the development tobacco use posit a central role of social relationships. Fundamental to these explanations is the well-researched tendency for individuals to be in relationships with peers who resemble them on a variety of behavioral and attitudinal characteristics (commonly referred to as homophily), and that peers facilitate the initiation and escalation of substance use (Nelson et al. 2015). In studies conducted in both the United States and Europe, tobacco use by peers predicted adolescents’ initial as well as continued use (Ali and Dwyer 2009). This has been well established with respect to tobacco use in the United States and Europe (e.g., Knecht et al. 2010).

Homophily occurs as a consequence of the synergistic processes of selection and influence (also referred to as socialization) (Prinstein and Dodge 2008). Selection describes a process by which youth establish friendships with others similar to themselves. Socialization occurs when friends influence each other in ways that lead them to change their behavior, perhaps contributing to the initiation of tobacco use as well as increases in rates of use (Mercken et al. 2009a). Most researchers have concluded that both selection and influence processes likely occur (Poulin et al. 2011), with some evidence that the effects of selection are stronger than those of socialization (Knecht et al. 2010; Mercken et al. 2009b). The influence processes that occur are likely to take the form of modeling and facilitation rather than overt pressure (Arnett 2007; Kobus 2002). The tendency for adolescents to resemble their peers in tobacco use appears to be fundamental to understanding adolescent tobacco use, the changes in such use over time, and may likely be relevant to constructing intervention programs.

Tobacco Use of Chinese Adolescents

There are multiple features of Chinese adolescents’ peer interactions that differ from those often present in the United States that could be potentially relevant to understanding the association between friendship and tobacco use. Chinese schools are organized by classrooms and adolescents typically stay within these groups throughout the day rather than move around from class to class as is more common in the United States. Because the membership of these classroom groups tends to remain constant from year to year, adolescents tend to be very familiar with those in their classroom, but less familiar with those in other classrooms. They generally have lower levels of romantic involvement than US adolescents (Li et al. 2010b). In addition, in comparison to US adolescents, Chinese students see themselves as less disrespectful, less oriented toward peers, more engaged in school, and less focused on attaining autonomy from the family (Qui et al. 2016). Chinese adolescents also spend less time associating with peers outside of school than students in the United States because they are more often at home studying (Stevenson and Zusho 2002), do not hold part time jobs, or drive cars. Thus, they often have less unsupervised time with friends than their counterparts in the United States and other countries and have fewer opportunity to interact with a wide range of peers in settings away from adult supervision during which they might use tobacco. Perhaps this could reduce the concordance between the tobacco use of adolescents and their school friends, but there is little research assessing this issue.

It is difficult to assess the rates of Chinese adolescent tobacco use because few comprehensive epidemiological studies have been conducted. The results from a 2004 study of adolescents in seven large Chinese cities (Ma et al. 2008) yielded lifetime rates of smoking of 33.3 % for boys and 15.8 % for girls—the percentage of adolescents who reported having smoked during the prior 30 days of 15.0 % for boys and 3.4 % for girls. The Global School-based Student Health Survey was administered in 2003 to about 9,000 11–15-year-old students in 4 Chinese cities and yielding rates of current smoking of 10.9 % for boys and 1.9 % for girls (Page et al. 2011). These high levels of adolescent male tobacco use parallels the high levels of tobacco use by Chinese adult men which is the second highest in Asia (Ott and Srinivsan 2012), with reports from a 2011Gallop survey that 57 % of men and 3 % of women smoke at least occasionally. This high level of tobacco use by Chinese adolescents has emerged as a major public health issue (Hesketh et al. 2001), and a target of intervention programs.

The two studies discussed above reveal large differences between the rates of tobacco use by Chinese boys and girls, patterns that deviate considerably from those seen in the United States. For example, results from Monitoring the Future (Miech et al. 2015) yielded roughly similar rates of prior month tobacco use for US tenth grade boys and girls, 7.7 % vs. 6.3 %, respectively. Perhaps these large differences in the smoking rates of boys and girls may reflect gender differences in the peer influence processes pertaining to tobacco use.

Peers and Chinese Adolescent Tobacco Use

Evidence from several studies suggests that peer relationships are associated with adolescent tobacco use in China. Jessor et al. (2003) found that the exhibition of problem behavior by adolescents’ peers predicted tobacco use in Chinese and US adolescent samples. Li et al. (2010a) found that peer tobacco use was associated with increased tobacco use among 15-year-old Chinese adolescents. Peer smoking was associated with an increase of 2.61 in the odds of smoking in a sample of 10–19-year-old Chinese adolescents (Zhang et al. 2000), and in both urban and rural China, adolescents whose friends smoked were two to four times more likely to have tried smoking themselves (Ma et al. 2008). In a large study of adolescents in several areas in China, smoking by best friends was associated with boys’ but not girls’ tobacco use (Grenard et al. 2006). Chen et al. (2006) similarly found that friends’ smoking has a stronger influence on the smoking of boys than girls. Thus, there is evidence that tobacco use by peers is associated with increased tobacco use of Chinese adolescents, although this effect may be more pronounced for boys than for girls.

Limitations of the Research Pertaining to Friends and Tobacco Use in China

It is difficult to interpret the findings from China regarding the similarity between the tobacco use of adolescents and their friends because of four methodological concerns that characterize much of the exiting research. First, researchers have often failed to be specific about the type of peer relationships under consideration, with insufficient appreciation that peers in these different types of relationships may vary with respect to their associations with adolescents’ tobacco use. Whereas some researchers have focused on friends, others have focused on peers in general. Second, because most of the information regarding both adolescents’ relationship with peers and peers’ tobacco use has come from self-reports, the existence of method error compromises conclusions about the extent to which adolescents are in fact similar to peers in their use of tobacco. Third, there has been limited longitudinal research assessing tobacco use by adolescents or their peers in China. Fourth, there has been little attention directed toward unraveling the influences of selection and influence in explaining the similarity between the tobacco use of adolescents and the friends. The existence of these methodological concerns make it difficult to determine the extent to which whether peer influence models of tobacco use developed in other countries are also appropriate for explaining the tobacco use of Chinese adolescents.

Diversity of Peer Relationships

Researchers have become increasingly cognizant of the diversity of child and adolescent peer relationships (Brechwald and Prinstein 2011; West and Michell, 1999). Adolescents interact not only with friends but with a wide variety of other peers, including acquaintances and youth whom they dislike. West and Michell (1999) concluded that friends appear to have the most influence on substance use. They cautioned, however, that it is difficult to interpret the research pertaining to this issue because many researchers have obtained information on peers’ substance use from adolescents’ reports regarding their peers rather than from the peers themselves. When researching the relation between peer relationships and tobacco, it is important to be specific about the type of relationship that is being studied. In the present study, we focused on the influence of mutual friends.

Friendships are optimally defined by mutual selection (Berndt and McCandless 2009) and by using this method, it is more likely that actual friendships rather than aspired friendships will be identified. It is particularly important to use mutual friendship reports because adolescents may not be accurate reporters of the existence of friendships, or their level of social acceptance (Brendgen et al. 2004; Feld and Carter 2002; Zakriski and Coie, 1996). US children appear to have difficulty estimating the extent to which they are liked by those in ethnic groups that differ from their own (Zakriski and Coie, 1996), aggressive children tend to over-estimate their social acceptance (e.g., Brendgen et al. 2004), and college students with low self-esteem have been found to be more accurate reporters of their social networks than others (Casciaro, 1998). These problems can be partially addressed by employing mutual rather than unilateral nominations of friendships.

Method Error

When adolescents report the tobacco use of both self and others, estimates of concordance between these are invariably confounded by the reporting biases that are shared across multiple variables. These biases include but are not limited to social desirability, tendencies to use or not use extreme scale values, and halo effects. Also contributing to the inflation of estimates of the association between ratings of self and others’ tobacco use is the false consensus effect, which refers to the tendency of individuals to incorrectly assume that others think and behave as they do. An empirical test of the effects of these potential biases was provided by Henry et al.'s (2011) analysis of two large data sets. Adolescents who did not use substances underestimated the extent to which their friends also refrained from use whereas those who used substances overestimated the extent to which their friends also did so. Similar results were obtained by Bellendiuk et al. (2010) who reported that adolescents are inaccurate reporters of the substance use of their peers. In sum, these biases may produce an inflated estimate of similarity between the substance use of adolescents and their friends (e.g., Iannotti and Bush, 1992; Kobus 2002; Prinstein and Wang 2005).

Longitudinal Assessment

The research on peer relationships and tobacco use of Chinese adolescents has been limited by the paucity of longitudinal research. In one of the few longitudinal studies of the effects of peers on the tobacco use of Chinese adolescents (Chen et al. 2006), best friend smoking at baseline predicted adolescent smoking 6 months later. In another study of smoking, Lin et al. (2008) assessed the behavioral change of friends over 6 months and found that the onset of friends’ smoking was associated with the onset of adolescents’ smoking. Because both adolescents’ and friend’s smoking were assessed from adolescents’ reports, the contribution from these studies is limited as these estimates of concordance may be inflated because of the potential of shared method error. Thus, there has been limited longitudinal study of tobacco use in China except for a few short-term longitudinal studies with methodological complications.

Selection and Influence

Our review of the prior studies of tobacco use Chinese adolescents and their peers reveals no studies focused on unraveling the processes of selection and influence. This is likely attributable in part to the paucity of longitudinal studies in China that could potentially address this question. Addressing this question is critically important for attempts to determine the applicability of peer influence models of tobacco use to Chinese adolescents.

The Present Study

This study was conducted to assess the relation between the tobacco use of adolescents and their friends in middle and high school while addressing the four methodological issues discussed above. First, we focused specifically on friendship as defined by mutual nominations. Second, we controlled for self-report biases by determining the identity of friends by verifying adolescents’ reports of these relationships from others, and obtaining reports of peer tobacco use directly from the peers themselves and not from the adolescents’ reports of the behavior of others. Third, we assessed the sample at three points, each separated by 1 year, thus affording the opportunity to use growth curve analyses to assess changes in tobacco use over time and to assess how changes in tobacco use of friends were associated with changes in adolescent tobacco use over this time period. Finally, cross-lagged analyses were used to assess selection and influence effects.

Method

Participants

Adolescents were recruited from urban schools in Lanzhou, Gansu, China, a provincial capital in Northwest China with a population of approximately 3.6 million. Almost all, 97 %, were of the majority Han nationality. There was a wide distribution of parent education levels; 42.9 % of mothers and 41.5 % of fathers had a junior high school education, 28.1 % of mothers and 32.6 % of fathers had a senior high school education, and 8.3 % of mothers and 11.9 % of fathers had a post-high school education. The large percentage of parents in this sample with less than a high school education is consistent with Chinese national statistics. The average years of parents’ schooling for 1986–1990 birth cohort was 10.4 years for urban children, according to the 2008 Rural-Urban Migration in China and Indonesia Survey (Golley and Kong 2013).

The middle school adolescents came from two schools with a total of 880 adolescents (399 girls) who participated in at least 1 year of data collection. The initial seventh grade sample included 766 adolescents (mean age = 13.33, sd = .64) in grade seven; of these 614 participated in the entire study. An additional 114 classmates of these original participants were recruited at grades eight and nine. This resulted in 764 participants in grade eight and 748 participants in grade nine.

The high school sample came from three high schools and included 849 adolescents (454 girls) who participated in at least one year of data collection. The initial tenth grade sample included 729 participants (mean age = 16.66, sd = .74); of these, 567 participated throughout the study. An additional 120 classmates were recruited at grades eleventh and twelve. This resulted in 783 participants in grade eleven and 662 in grade twelve.

Attrition analyses were conducted comparing those who remained in the study during the entire 3 years with those who entered in Year 1 and left in either Year 2 or Year 3. Independent t-tests revealed that adolescents who left the study did not differ in their level of tobacco use or parental education; their friends, however, smoked more than the friends of those who remained in the study (d = .21).

Measures

Mutual Friendships

Participants identified a maximum of five friends. They were presented with a list of classmates and selected from this list or provided the names of their male and female friends who attended their school and were in the same grade. The average number of mutual friendships ranged from 1.90 to 2.12 from seventh to ninth grade, and from 2.22 to 2.35 from tenth to twelfth grade. Most of the friendships were between members of the same sex, 91 % to 97 % in middle school and 90 % to 92 % in high school. In middle school there were 98 students at grade seven, 134 students at grade eight, and 90 students at grade nine who did not have a mutual friend. In high school, there were 52 students at grade ten, 80 students at grade eleven, and 81 at grade twelve who did not have mutual friends. There were no significant differences in the use of tobacco between those with and without friends. Consistently across grades, boys more often than girls did not have mutual friends (χ2 ranged from 11.85 to 26.25, p < .001).

Tobacco Use of Adolescents and Their Friends

Participants rated their tobacco use during the prior year using a 4-point ordinal scale: 0 = never, 1 = once or twice, 2 = monthly use (once or twice a month), 3 = weekly use (more than once a week). Table 1 presents frequency of boys’ and girls’ tobacco use in middle and high school.
Table 1

Frequency of boys’ and girls’ tobacco use in middle and high school

 

Middle school

High School

 

Boys

Girls

Boys

Girls

Frequency

Percentage

Frequency

Percentage

Frequency

Percentage

Frequency

Percentage

Year 1 (7th/10th)

      

0

304

75.1

335

94.6

173

53.7

372

92.1

1

62

15.3

15

4.2

35

10.9

21

5.2

2

23

5.7

4

1.1

34

10.6

9

2.2

Year 2 (8th/11th)

      

0

276

66.8

322

92.3

201

56.6

411

96.3

1

39

9.4

14

4.0

30

8.5

6

1.4

2

36

8.7

8

2.3

24

6.8

9

2.1

3

62

15.0

5

1.4

100

28.2

1

0.2

Year 3 (9th/12th)

      

0

240

61.1

327

93.2

173

58.2

347

95.3

1

43

10.9

18

5.1

21

7.1

7

1.9

2

17

4.3

4

1.1

15

5.1

4

1.1

3

93

23.7

2

0.6

88

29.6

6

1.6

Note: 0 = never use; 1 = once or twice in the past year; 2 = monthly use; 3 = weekly use

Procedure

The procedures and consent and assent procedures were overseen by the Purdue University IRB. Parents were provided with information about the study and returned signed consent forms; adolescents were provided with a description of the study and provided written assent to participate. Assessments were conducted in classroom groups under the supervision of Psychology graduate and undergraduate students. Average classroom participation was 89 % and ranged from 57 % to 100 %.

Results

Table 2 presents descriptive statistics of adolescents’ and friends’ smoking during middle and high school. The missing data for friends’ tobacco use is explained by the absence of data for adolescents who did not have a mutual friend at a given assessment period. Ninety-three percent of middle school adolescents and 96 % of high school students had mutual friends’ reports for two or more assessment periods; ten middle school students and five high school students had no mutual friends at any time.
Table 2

Means and standard deviations of study variables

 

Middle school

High School

 

Boys

Girls

Boys

Girls

 

m

sd

m

sd

m

sd

m

sd

Tobacco Y1

.39

.77

.06

.29

1.07

1.28

.11

.42

Tobacco Y2

.72

1.13

.13

.49

1.06

1.33

.06

.34

Tobacco Y3

.91

1.26

.09

.37

1.06

1.35

.09

.45

Friend’s Tobacco Y1

.40

.63

.08

.23

.97

1.01

.17

.43

Friend’s Tobacco Y2

.66

.89

.16

.41

1.02

1.10

.11

.32

Friend’s Tobacco Y3

.79

1.03

.20

.47

.85

.99

.10

.35

Table 3 presents correlations between boys’ and girls’ smoking with their friends’ smoking across the 3 years. Boys’ tobacco use was significantly and positively correlated with their friends’ use in both middle and high schools (r’s ranged from .32 to .48). In contrast, girls’ tobacco use was significantly correlated with their friends’ use only at eighth, tenth, eleventh, and twelfth grade (r’s ranged from .12 to .30). Comparisons of independent correlation coefficients (Cohen and Cohen, 1983) revealed that correlations between girls and their friends were lower than those of boys at all grades with the exception of grade twelve (z’s ranged from 2.45 to 4.44).
Table 3

Correlations between boys’ and girls’ smoking, and mutual friends’ smoking in middle and high schools

 

1

2

3

4

5

6

Middle school

 1. Smoke Y1

.32***

.17**

.10

.04

.09

 2. Smoke Y2

.55***

.54***

−.04

.13*

.18**

 3. Smoke Y3

.42***

.57***

−.02

.03

.01

 4. Friend smoke Y1

.41***

.29***

.24***

−.01

.10

 5. Friend smoke Y2

.37***

.32***

.22***

.38***

.25***

 6. Friend smoke Y3

.16**

.30***

.41***

.28***

.35***

High school

 1. Smoke Y1

.37***

.55***

.23***

.11*

.02

 2. Smoke Y2

.79***

.57***

−.01

.12*

.13*

 3. Smoke Y3

.71***

.78***

.06

.12*

.30***

 4. Friend smoke Y1

.48***

.47***

.36***

.26***

.24***

 5. Friend smoke Y2

.40***

.45***

.36***

.56***

.20***

 6. Friend Smoke Y3

.30***

.37***

.37***

.29***

.42***

Note: Correlations for boys are below the diagonal and those for girls are above the diagonal. Correlations between adolescents and their friends are bolded

p < .10; *p < .05; **p < .01; ***p < .001

Linear Growth Models

Our attempts to apply linear growth curve models to the longitudinal data were successful for boys but not for girls, This is likely attributable to the very low frequency of tobacco use by girls in conjunction with their tobacco use not conforming to a clear developmental trajectory. In middle school (Table 1), there were 5 % of girls who had ever smoked (compared to 25 % of boys) at seventh grade and this increased slightly to 7 % at ninth grade. In high school, about 8 % of girls had ever smoked at tenth grade (compared to 46 % of boys) and that number dropped to about 4 % at eleventh and twelfth grade. Thus, in contrast to boys, there was no clear trajectory of either stable or increasing tobacco use over time.

Linear growth models of tobacco use were estimated separately for boys and their friends. Because moderation analyses revealed that the intercepts and slopes differed for middle and high school, separate growth models were developed for these two cohorts. Trajectory mean and variance estimates are presented in Table 4.
Table 4

Unstandardized intercept and slope estimates for the trajectories of boy’s’ tobacco use and friends’ tobacco use

 

Middle school

High school

 

Adolescent

Friends

Adolescent

Friends

Indices

Estimate

SE

Estimate

SE

Estimate

SE

Estimate

SE

Averages

 Intercept

.41***

.04

.39***

.04

1.04***

.07

.97***

.06

 Slope

.28***

.03

.22***

.03

.04

.03

−.03

.04

Individual differences (variance)

 Intercept

.49***

.08

.26***

.07

1.41***

.15

.92***

.14

 Slope

.19***

.05

.11*

.04

.15*

.06

.22***

.06

 Covariance (Intercept/Slope)

.00

.05

−.04

.04

−.10

.07

−.30***

.08

Note: *p < .05; **p < .01; ***p < .001

The linear model for middle school boys’ smoking had a good fit, χ2 (1) = .68, ns; CFI = 1.00; RMSEA = .00, and revealed a significant increase in boys’ tobacco use from seventh to ninth grade. In addition, statistically significant individual differences were present at both baseline and in the rate of change over time. A separate linear model for friends’ tobacco use also fit the data well, χ2 (1) = 1.65, ns; CFI = .99; RMSEA = .04. Friends’ tobacco use also had a significant increase over time with significant individual variation in both intercepts and slopes.

The model for high school boys’ tobacco use also had a good fit, χ2 (1) = .79, ns; CFI = 1.00; RMSEA = .00. Unlike the increases in tobacco use over time observed in middle school, tobacco use remained relatively stable over the 3 years in high school as evidenced by the nonsignificant slope. There were, however, significant individual differences in both baseline smoking and the rate of change over time. Similarly, a separate model for friends’ tobacco use had a good fit, χ2 (1) = 2.46, ns; CFI = .99; RMSEA = .06, with no significant slope effects, but with significant individual differences in baseline and slope.

Parallel Process Models

The trajectories for boys’ and their friends’ tobacco use for middle and high school were simultaneously estimated using parallel process models. The middle school model had a good fit, χ2 (7) = 12.76, ns; CFI = .99; RMSEA = .04. Positive associations between adolescents’ and friends’ tobacco use emerged for both intercepts, r = .68, p < .001, and slopes, r = .69, p < .001. The high school model also had a good fit, χ2 (7) = 9.30, ns; CFI = .99; RMSEA = .03. Positive associations were found between adolescents’ and friends’ tobacco use for both intercepts, r = .56, p < .001, and slopes, r = .39, p < .01.

Cross-Lagged Models

Cross-lagged models linking the smoking of boys and their friends during middle and high school are presented in Figs. 1 and 2; comparable models for girls did not meet minimally acceptable model fit criteria. In the middle school sample, the model fit the data well, χ2 (5) = 5.86, ns; CFI = .99; RMSEA = .02. Boys’ smoking consistently predicted friend’s smoking the following year; however, smoking by friends did not predict the subsequent tobacco use of boys. The high school model also fit the data well, χ2 (5) = 5.13, ns; CFI = 1.00; RMSEA = .00, with the result that boys’ smoking predicted subsequent changes in friends’ smoking from tenth to eleventh grades and from eleventh to twelfth grades. In contrast, friend’s smoking predicted changes in boys’ smoking only from tenth to eleventh grade. We assessed the relative strength of these two tenth to eleventh grade pathways by constraining them to be equal with the result that the two pathways were not statistically different (∆χ2 = .05, ns.).
Fig. 1

Cross-lagged model for adolescent boys’ and friends’ tobacco use in middle school. Standardized parameter estimates were presented. All parameter estimates were significant at .001

Fig. 2

Cross-lagged model for adolescent boys’ and friends’ tobacco use in high school. Standardized parameter estimates were presented. All parameter estimates were significant at .001, except for the one in parenthesis which was significant at .01

Discussion

The concordance between the smoking of adolescents and their friends is central to peer influence models of tobacco use (Brechwald and Prinstein 2011). This study of Chinese adolescents was initiated to assess the extent to which the relationship between the tobacco use of Chinese adolescents is associated with the smoking of mutual friends in a manner similar to what has emerged from studies in the United States and Europe. Of particular interest was our efforts using path analyses to explore the extent to which similarities stem from selection, influence, or the combination of these. In doing so, we focused on several questions, including potential methodological confounds, gender differences, and the longitudinal relations between the tobacco use of adolescents and their friends. Because many features of Chinese adolescence differ from those typical in the United States, the study of peer influence in China provides an excellent test of the generalizability of peer influence models.

An important feature of the present study was our effort to address several methodological problems that have characterized much of the past research on tobacco use in China. Almost all of the research that has focused on peers and tobacco use in China has exclusively utilized self-report data. The presence of reporting biases are relevant to both identification of friends and the assessment of tobacco use of friends. To address the possibility of reporting biases regarding the existence of relationships, friendships were identified only if both parties agreed upon their existence. Information about friends’ tobacco use was obtained using reports from these individuals rather than relying upon reports of the participants regarding the use of others. Not only are such reports likely to be inaccurate given the lack of knowledge of about others, but they may also be systematically biased as a consequence of shared method effects. In addition, the reports of the smoking of self and the friend are likely inflated due to the existence of the false consensus effect, in which individuals are likely incorrectly to believe that others use substances at rates similar to themselves (Henry et al. 2011; Iannotti and Bush, 1992; Prinstein and Wang 2005). This overestimation of peer smoking has been reported to be common among Chinese adolescents in Hong Kong (Lai et al. 2004). Thus, our estimates of the concordance between the smoking of adolescents and their friends control for these potential confounds.

Rates of Tobacco Use

Based on prior finding from China, it was anticipated that boys’ tobacco use would be greater than that reported from US adolescents, but that girls’ smoking rates would be lower than that reported for US girls. There were 23.7 % eighth grade boys and 35.4 % tenth grade boys who were reported using tobacco once a month or more. These rates can be compared with the respective rates of 3.3 % and 6.1 % derived from the Monitoring the Future reports of 2015 US adolescent tobacco use (Johnston et al. 2016). The rates for Chinese girls monthly tobacco use are similar to that of US girls (3.7 % for both counties), but lower than that in tenth grade (2.7 % vs. 6.3 %).

The results from the trajectory analysis revealed that boys increased their rates of tobacco use from seventh to tenth grade, but that these trajectories were essentially flat during high school. These findings are consistent with the results from the China Seven Cities Study (Sun et al. 2006) that the transitions into smoking were more rapid for middle school than for older adolescents. Although these developmental patterns appear generally consistent with findings from the United States (Nelson, et al. 2015), further large population longitudinal research is needed to more specifically determine the similarities and differences between the trajectories of tobacco use in the two countries.

Associations between Tobacco Use of Adolescents and Their Friends

Our primary goal was to understand the extent to which the substance use by friends was associated with adolescent substance use. The results differed for boys and girls. Inspection of the correlations between the tobacco use of boys and their friends reveals consistent patterns of associations across the six grades with correlations that ranged between .32 and .48 between adolescents and their friends. The results from the parallel process model add to these findings by showing significant associations between both the intercepts and slopes for boys and friends smoking in both middle and high school. Finally, there were significant pathways linking the smoking of boys and that of their friends in the cross-lagged models. These results provide support to the arguments of West and Michell (1999) that there are strong associations between the tobacco use of adolescents and their friends, at least for Chinese boys.

The associations between the tobacco use of girls and their friends were less consistent than those found for boys. The correlations between the smoking of girls and their friends were significant in only four of the six grades, and these effects were lower in magnitude than those of boys in five of the six grades. It is possible that these gender differences reflect differences in the peer influence processes pertaining to boys’ and girls’ tobacco use.

Longitudinal Pathways: Selection and Influence

The longitudinal design afforded an exploration of the cross-lagged effects linking the smoking of boys and their friends. The results revealed consistent findings of pathways from boys’ smoking to that of their friends the subsequent year. In contrast, only one of the four pathways from friends’ tobacco use to boys, tobacco use was found (i.e., tenth to eleventh grade). Thus, although some evidence of bidirectional effects emerged, selection effects more consistently emerged.

Considerable research is necessary to further untangle how selection and influence processes contribute to the observed similarity between the tobacco use of Chinese adolescents and their friends. We suspect that those who smoke tobacco are attracted to others who also do so. But we do not know whether shared interests in smoking brings these adolescents together, or whether their companionship is driven by other shared characteristics such engagement in deviant behavior, popular status, or low academic achievement. Further research on the relations between tobacco use and other aspects of adjustment and competence are needed. This research will be particularly useful for understanding the smoking of girls.

It is also likely that relationships developed on the basis of shared interests in smoking or other activities or shared characteristics are also sources of influence. For example, there are likely processes of shaping whereby friends influence each other to engage in more extreme substance use and other forms of deviance (Dishion et al. 2008). Being friends with someone who uses tobacco might also facilitate increased use of tobacco as their social interactions might be occasions of engaging in this behavior.

There are multiple limitations regarding the value of these path analyses as means to unravel the effects of selection and influence. First, aggregate rather than individual friendships were used in these analyses. Thus, the friendship scores do not reflect individual friends, but instead that of the smoking of an assortment of between one and five friends; consequently, selection and influence processes as these occur within single friendships might be hidden. Second, although the use of reciprocated friendships is advantageous because we have confidence that these relationships are real and not aspirational it is important to attend to unilateral friendships as these reflect the reaching out to others as potential friends (Brechwald and Prinstein 2011). Third, friendships do not exist in isolation, but are instead nested with networks of interconnected friendship links. Thus, it is important to consider the overall network in addition to the dyadic friendships assessed in the present study. There are two approaches that can be used to further unravel selection and influence effects.

To gain further insight into influence and selection processes, it will be valuable in future studies to conduct assessments more frequently than the yearly assessment schedule employed in this study. Poulin et al. (2011) conducted bi-monthly assessments of friendship and substance use of US ninth grade adolescents, and found that friendships were unstable over this period and that changes in friendship were associated with changes in adolescent substance use, and reflected both selection and influence effects. It will be useful to employ similar procedures to assess the friendships and tobacco use of Chinese adolescents. Such study may be particularly illuminative with respect to understanding peer influence processes for girls. In the present study, girls’ tobacco use did not fit a growth curve model and there were low and inconsistent associations between the tobacco use of girls and their friends. Perhaps knowledge of the specific changes in friendship and how these are associated with changes in tobacco use will clarify peer influence processes as these pertain to girls’ tobacco use. Of particular interest will be understanding girls’ relationships with boys either as friends or romantic partners, given findings that such associations are particularly salient for US girls’ tobacco use (Mrug et al. 2011).

Second, an important recent advance in the attempts to tease apart the influence of selection and influence is the development the Simulation Investigation for Empirical Network Analysis (SIENA) (Snijders et al. 2010; Steglich et al. 2006). SIENA estimates whether similarity longitudinally emerges from peer selection, peer influence, or a combination of both (Snijders et al. 2007; Steglich et al. 2010; Veenstra et al. 2013). This anlytic strategy makes use of unilateral nominations and incorporates information about the friendship network in addition to specific friendship ties. It will be useful in further studies to assess the convergence between SIENA results and those of the longitudinal analyses path analyses used in this study.

Gender Differences in Chinese Adolescent Tobacco Use

Our finding of large gender differences in the rates of smoking parallels those found in prior studies of Chinese adolescents. Although the processes that underlie gender differences in the Chinese adolescent smoking have been only minimally explored, these different rates of use suggests the possibility that there are differences between boys and girls in the reasons for smoking, the characteristics of boys and girls who smoke, and the attributions that others make about boys and girls who smoke. Further quantitative and qualitative research is necessary to explore these and other topics.

The present findings provide evidence of gender differences in the magnitude of the association between the smoking of adolescents and their friends. Whereas there is consistent evidence across grades and analyses that there are strong association between the smoking of boys and their friends, the findings for girls were less consistent and the correlations were lower in magnitude than those for boys. These findings are consistent with the results from Chen et al. (2006) and Grenard et al. (2006) in suggesting that the relation between adolescents and friends smoking is larger for boys than for girls in China. Although this finding appears to be robust, there is little information about why this is the case. One possibility is that girls’ smoking is more strongly influenced by a few select members of their friendship group rather than the aggregate. They may also be influenced by their non-friend companions, perhaps by members of their larger social network or romantic partners (Kobus 2002). The present finding, however, provide further evidence that the dynamics underlying Chinese boys’ and girls’ tobacco use differs and that further research is required to understand this.

Qualitative research is needed to explore reasons why boys and girls smoke, variation in attitudes toward smoking, attitudes about smoking and smokers, and the physical and relational contexts in which smoking occurs. Okamota et al. (2012) conducted focus groups exploring these topics with high school adolescents from southern China. Emerging from these groups were suggestions that girls’ smoking was seen negatively and that this behavior reflected badly upon girls’ but not boys’ character. A minority position, however, was that girls’ smoking was modern and fashionable. Further research of this type is needed to determine the factors that underlie these gender differences in tobacco use. Given the increased independence of contemporary youth and their exposure to international influences, it is important to determine whether traditional negative views of smoking by girls are changing. There is a widespread belief that the usage of tobacco use of girls has increased in recent years (Wen et al. 2007), but the requisite research necessary to assess this has not been done.

Limitations

Although this study assessed adolescents in grades seven through twelve, an important limitation is that we assessed two separate samples, one in middle school and the other in high school. Consequently, our ability to examine the trajectories of tobacco across the middle to high school transition is limited. Nevertheless, our results appear consistent with findings of Bernat et al. (2008) that adolescent tobacco use of Chinese adolescents is often initiated prior to 14 years of age.

Because this research was conducted only in China, it is impossible to make direct cross-cultural comparisons of the processes associated with tobacco use. Although we can compare our results with findings obtained from studies of adolescents in the United States and other countries, we cannot directly test these comparisons. It would be particularly useful to institute a large scale assessment of the substance use of Chinese adolescents using procedures employed in the Monitoring the Future studies (Johnston et al. 2016) as this would enable direct comparisons across countries.

Our analyses were limited to adolescents’ same-age friends in school and we do not know whether they use tobacco with non-school friends in other contexts. Perhaps tobacco use is more likely a function of adolescents’ associations with neighborhood friends than with classmates. It is also possible that adolescents associate with others who are older or younger than themselves and that these associations have implications for tobacco use. Unfortunately, our knowledge of Chinese adolescent peer relationships is very limited; the scattered reports that exist tend to focus on school contexts with little known about peer relationships in other contexts.

Any discussion of tobacco use in China must reflect the rapid societal changes that have occurred and their effects on adolescent development (Liu et al. 2012). Chinese adolescents are becoming more individualistic and independent from parents. It is unknown how these changes have impacted tobacco use, and whether these have more of an impact on the gender variation in tobacco use.

The participants in this study came from typical academic schools that served a diverse SES population located in a provincial capital in central China. Many Chinese students attend vocational schools, drop out prior to the twelfth grade, or enroll in elite academic schools. It is likely that the rates of tobacco adolescents in these differ from the rates found in this study. It should also be considered that there are vast regional differences in China and that the rates of tobacco use likely differ as a function of the diverse geographic areas of China, economic status, and urban vs. rural residency.

Conclusion

Central to theories of peer influence is the question of the similarity between adolescents and their friends in tobacco use over time. The present findings are consistent with past findings that Chinese boys use tobacco at substantially higher rates than girls. Boys’ tobacco use fits a linear growth model with escalating use through middle school with a flattening of the trajectory in high school, whereas girls’ smoking did not appear to fit into a linear trajectory model. Boys’ tobacco use was consistently and strongly associated with that of their friends both concurrently and longitudinally. In contrast, the associations between the tobacco use of girls and their friends were inconsistent and lower in magnitude that those of boys. This suggests that possibility that peer influence processes for Chinese boys and girls differ, but to explore this, further quantitative and qualitative studies of this are needed. Finally, attempts to partial out selection and influence processes have not been done in China with respect to tobacco use. Consistent evidence of selection effects emerged for boys, suggesting that they became friends with those who resembled them in levels of tobacco use. We were unable, however, to examine this in girls because the longitudinal models fit poorly. Our findings reveal gender differences in tobacco use, but further research is needed to understand the peer influence processes that might explain this.

Further research elucidating peer influence processes that pertain to tobacco use of Chinese adolescents will be particularly useful for understanding the extent to which models of peer influence developed in Western countries are widely generalizable. Such research, however, needs to extend beyond attempts to unravel selection and influence attempts to explore the mechanisms by which peer influence occurs (Brechwald and Prinstein 2011; Kobus 2002). Important questions that require further study include differential moderators of peer influence that may include characteristics of the individual being influenced and those of the influencer in addition to features of the relationship, such as whether the individuals currently have a relationship or whether an individual desires a relationship with the other. Only beginning efforts have been made to understand culture and peer influence with some evidence that the impact of peers varies across cultures (Astudillo et al. 2013), as well as evidence that the relative influence of selection and influence processes pertaining to tobacco use varies across countries (Mercken et al. 2009b). It is also possible that cultural variation exists in the extent to which peer influence processes are similar for boys and girls (Okoli et al. 2013). Considerable research is required to understand culture and peer influence pertaining to tobacco use; systematic research in China will make a substantial contribution toward this effort.

Notes

Acknowledgments

The authors appreciate insights of Handrea Logis pertaining to the discussion of selection and influence in this manuscript

Author Contributions

: LI and SJ were involved in the initial design of the study, were completely responsible for the data collection, conducted analyses, and reviewed articles on tobacco use written in Chinese journals; LI, a doctoral student, took the lead in this effort and was supervised by SJ, her major professor. TL conducted the path analyses and participated in the writing of the publication. LN and YF managed the data set from the US side, conducted some of the analyses, and participated in the literature research and writing of the study. DF coordinated the research and was involved in all parts of the project including design, analysis, and presentation. All authors read and approved the final manuscript.

Funding

This research was supported by a Major Project of National Social Science Foundation of China. This grant is identified as (13&ZD073).

Compliance with Ethical Standards

All procedures used in this study were consistent with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Conflict of interest

The authors declare that they have no competing interests.

Ethical Approval

The conduct of this research was approved by Purdue University Institutional Review Board.

Informed Consent

Signed parent consent and signed adolescent assent were obtained from all participants.

References

  1. Ali, M. M., and Dwyer, D. S. (2009). Estimating peer effects in adolescent smoking behavior: A longitudinal analysis. Journal of Adolescent Health, 45, 402–408. doi:10.1016/j.jadohealth.2009.02.004 CrossRefPubMedGoogle Scholar
  2. Arnett, J. J. (2007). The myth of peer influence in adolescent smoking initiation. Health Education and Behavior, 34, 594–607. doi:10.1177/1090198105285330 CrossRefPubMedGoogle Scholar
  3. Astudillo, M., Conner, J., Roiblatt, R. E., Ibanga, A. K. J., and Gmel, G. (2013). Influence from friends to drink more or less: A cross-national comparison. Addictive Behaviors, 38, 2675–2682. doi:10.1016/j.addbeh.2013.06.005 CrossRefPubMedGoogle Scholar
  4. Bellendiuk, K. A., Molina, B. S. G., and Donovan, J. E. (2010). Concordance of adolescent reports of friend alcohol use, smoking, and deviant behavior as predicted by quality of relationships and demographic variables. Journal of Studies on Alcohol and Drugs, 71, 253–257. doi:10.15288/jsad.2010.71.253 CrossRefGoogle Scholar
  5. Bernat, D. H., Erickson, D. J., Widome, R., Perry, C. L., and Foster, J. L. (2008). Adolescent smoking trajectories: Results form a population-based cohort study. Journal of Adolescent Health, 43, 334–340.CrossRefPubMedPubMedCentralGoogle Scholar
  6. Berndt, T. J., and McCandless, M. A. (2009). Methods for investigating children’s relationships with friends. In K. H. Rubin, W. M. Bukowski and B. Laursen (Eds.), Handbook of peer interactions, relationships, and groups (pp. 63–81). New York: Guilford Press.Google Scholar
  7. Brechwald, W.A., and Prinstein, M.J. (2011). Beyond homophily: A decade of advances in understanding peer influence processes. Journal of Research on Adolescence, 21, 166–179. doi:10.1111/j.1532-7795.2010.00721.x CrossRefPubMedPubMedCentralGoogle Scholar
  8. Brendgen, M., Vitaro, F., Turgeon, L., Poulin, F., and Wanner, B. (2004). Is there a dark side of positive illusions? Overestimation of social competence and subsequent adjustment in aggressive and nonaggressive children. Journal of Abnormal Child Psychology, 32, 305–320. doi:10.1023/b:jacp.0000026144.08470.cd CrossRefPubMedGoogle Scholar
  9. Casciaro, T. (1998). Seeing things clearly: Social structure, personality, and accuracy in social network perception. Social Networks, 20, 331–351. doi:10.1016/S0378-8733(98)00008-2 CrossRefGoogle Scholar
  10. Chen, X., et al. (2006). Perceived smoking norms, socioenvironmental factors, personal attitudes, and adolescent smoking in China: A mediation analysis with longitudinal data. Journal of Adolescent Health, 38, 359–368. doi:10.1016/j.jadohealth.2005.03.010 CrossRefPubMedGoogle Scholar
  11. Cohen, J., and Cohen, P. (1983). Applied multiple regression/correlation analysis for the behavioral sciences. Hillsdale, NJ: Erlbaum.Google Scholar
  12. Dishion, T. J., Piehler, T. F., and Myers, M. W. (2008). Dynamics and ecology of adolescent peer influence. In M. J. Prinstein and K. A. Dodge (Eds.) (2008). Understanding peer influence in children and adolescents (pp. 72–93). New York: Guilford Press.Google Scholar
  13. Feld, S. L., and Carter, W. C. (2002). Detecting measurement bias in respondent reports of personal networks. Social Networks, 24, 365–383.CrossRefGoogle Scholar
  14. Golley, J., and Kong, S. T. (2013). Inequality in intergenerational mobility of education in China. China and World Economy, 21, 15–37.CrossRefGoogle Scholar
  15. Grenard, J. L., Guo, Q., Jasuja, G. K., Unger, J. B., Chou, C., Gallaher, P. E., and Johnson, C. A. (2006). Influences affecting adolescent smoking behavior in China. Nicotine and Tobacco Research, 8, 245–255. doi:10.1080/14622200600576610 CrossRefPubMedGoogle Scholar
  16. Henry, D. B., Kobus, K., and Schoeny, M. E. (2011). Accuracy and bias in adolescent’s perceptions of friends’ substance use. Psychology of Addictive Behaviors, 25, 80–89. doi:10.1037/a0021874 CrossRefPubMedPubMedCentralGoogle Scholar
  17. Hesketh, T., Ding, Q. J., and Tomkins, A. (2001). Smoking among youths in China. American Journal of Public Health, 91, 1653–1655. doi:10.2105/AJPH.91.10.1653 CrossRefPubMedPubMedCentralGoogle Scholar
  18. Iannotti, R., and Bush, P. J. (1992). Perceived vs. actual friends’ use of cigarettes, marijuana, and cocaine: Which has the most influence?. Journal of Youth and Adolescence, 21, 375–389. doi:10.1007/BF01537024 CrossRefPubMedGoogle Scholar
  19. Jessor, R., Turbin, M. S., Costa, F. M., Dong, Q., Zhang, and Wang, C. (2003). Adolescent problem behavior in China and the United States: A cross-national study of psychosocial protective factors. Journal of Research on Adolescence, 13, 329–360. doi:10.1111/1532-7795.1303004 CrossRefGoogle Scholar
  20. Johnston, L. D., O’Malley, P. M., Miech, R. A., Bachman, J. G., and Schulenberg, J.E. (2016). Monitoring the Future national survey results on drug use, 1975-2015: Overview, key findings on adolescent drug use. Ann Arbor, MI: Institute for Social Research, The University of Michigan.Google Scholar
  21. Knecht, A. B., Burk, W. J., Weesie, J., and Steglich, C. (2010). Friendship and alcohol use in early adolescence: A multilevel social network approach. Journal of Research on Adolescence, 21, 475–487. doi:10.1111/j.1532-7795.2010.00685.x CrossRefGoogle Scholar
  22. Kobus, K. (2002). Peers and adolescent smoking. Addiction, 98, 37–55.CrossRefGoogle Scholar
  23. Lai, M. K., Ho, S. Y., and Lam, T. H. (2004). Perceived peer smoking prevalence and its association with smoking behaviors and intentions in Hong Kong Chinese adolescents. Addiction, 99(9), 1195–1205.CrossRefPubMedGoogle Scholar
  24. Li, X., Mao, R., Stanton, B., and Zhao, Q. (2010a). Parental, behavioral, and psychological factors associated with cigarette smoking among secondary students in Nanjing, China. Journal of Child and Family Studies, 19, 308–317. doi:10.1007/s10826-009-9299-1 CrossRefGoogle Scholar
  25. Li, Z. H., Connolly, J., Jiang, D., Pepler, D., and Craig, W. (2010b). Adolescent romantic relationships in China and Canada: A cross-national comparison. International Journal of Behavioral Development, 34(2), 113–120.CrossRefGoogle Scholar
  26. Lin, D., Fang, X., and Li, X. (2008). Huang-jing-he-ge-ti-yin-su-yu-qing-shao-nian-xi-yan-xing-wei-de-fa-sheng [The relationship of environmental and individual factors with adolescent smoking onset]. Xin-Li-Ke-Xue, 31(2), 304–307. doi:10.3969/j.issn.1671-6981.2008.02.010 Google Scholar
  27. Liu, J., Li, D., Chen, X., and French, D. C. (2012). Shyness-sensitivity, aggression, and adjustment in urban Chinese adolescents at different historical times. Journal of Research on Adolescence, 22, 393–399.CrossRefGoogle Scholar
  28. Ma, H., et al. (2008). Risk factors for adolescent smoking in urban and rural China: Findings from the China seven cities study. Addictive Behaviors, 33, 1081–1085. doi:10.1016/j.addbeh.2008.04.004 CrossRefPubMedGoogle Scholar
  29. Mercken, L., Candel, M., Willems, P., and deVries, H. (2009a). Social influence and selection effects in the context of smoking behavior: Changes during early and mid-adolescence. Health Psychology, 28, 73–82. doi:10.1037/a0012791 CrossRefPubMedGoogle Scholar
  30. Mercken, L., Snijders, T. A. B., Steglich, C., Vartiainen, E., and deVries, H. (2009b). Dynamics of adolescent friendship networks and smoking behavior: Social network analyses in six European countries. Social Science and Medicine, 69, 1506–1514. doi:10.1016/j.socscimed.2009.08.003 CrossRefPubMedGoogle Scholar
  31. Miech, R. A., Johnston, L. D., O’Malley, P. M., Bachman, J. G., and Schulenberg, J. E. (2015). New trends in teen smoking, e-cigarettes in 2015. Ann Arbor, MI: University of Michigan News Service. http://www.monitoringthefuture.org. Accessed 20 May 2016.
  32. Mrug, S., Borch, C., and Cillessen, A. H. N. (2011). Other-sex friendships in late adolescence: Risky associations for substance use and sexual debut? Journal of Youth and Adolescence, 40, 875–888. doi:10.1007/s10964-010-9605-7 CrossRefPubMedGoogle Scholar
  33. Nelson, S. E., Van Ryzin, M. J., and Dishion, T. J. (2015). Alcohol, marijuana, and tobacco use trajectories from age 12 to 24 years: Demographic correlates and young adult substance use problems. Development and Psychopathology, 27, 253–277. doi:10.1017/S0945/9414000650 CrossRefPubMedGoogle Scholar
  34. Okamota, J., Sakuma, K., Yan, H., Qiu, P., Palmer, P.H., and Johnson, C.A. (2012). A qualitative exploration of youth in the “new” China: Perspectives on tobacco use from adolescents in southwest China. Asia-Pacific Journal of Public Health, 24, 296–307.CrossRefGoogle Scholar
  35. Okoli, C., Greaves, L., and Fagyas, V. (2013). Sex differences in smoking initiation among children and adolescents. Public Health, 127, 3–10. doi:10.1016/j.puhe.2012.09.015 CrossRefPubMedGoogle Scholar
  36. Ott, B., and Srinivsan, R. (2012). Three in 10 Chinese adults smoke. http://www.gallup.com/poll/152546/three-chinese-adults-smoke.aspx. Accessed 12 Oct 2014.
  37. Page, R. M., Dennis, M., Lindsay, G. B., and Merrill, R. M. (2011). Psychosocial distress and substance use among adolescents in four countries: Philippines, China, Chile, and Namibia. Youth and Society, 43, 900–930. doi:10.1177/0044118X10368932 CrossRefGoogle Scholar
  38. Poulin, F., Kiesner, J., Pedersen, S., and Dishion, T. J. (2011). A short-term longitudinal analysis of friendship selection on early adolescent substance use. Journal of Adolescence, 34, 249–256. doi:10.1016/j.adolescence.2010.05.006 CrossRefPubMedGoogle Scholar
  39. Prinstein, M. J., and Dodge, K. A. (Eds.). (2008). Understanding peer influence in children and adolescents. New York: Guilford Press.Google Scholar
  40. Prinstein, M. J., and Wang, S. S. (2005). False consensus and adolescent peer contagion: Examining discrepancies between perceptions and actual reported levels of friends’ deviant and health risk behaviors. Journal of Abnormal Child Psychology, 33, 293–306. doi:10.1007/s10802-005-3566-4 CrossRefPubMedGoogle Scholar
  41. Qui, Y., Pomerantz, E. M., Wang, M., Cheung, C., and Cimpian, A. (2016). Conceptions of adolescence: Implications for differences in engagement in school over early adolescence in the United States and China. Journal of Youth and Adolescence, 45, 1512–1526. doi:10.1007/s10964-016-0492-4 CrossRefGoogle Scholar
  42. Snijders, T. A. B., van de Bunt, G. G., and Steglich, C. E. G. (2010). Introduction to actor-based models for network dynamics. Social Networks, 32, 44–60.CrossRefGoogle Scholar
  43. Snijders, T. A. B., Steglich, C. E. G., and Schweinberger, M. (2007). Modeling the co-evolution of networks and behavior. In K. van Montfort, H. Oud and A. Satorra (Eds.), Longitudinal models in the behavioral and related sciences (pp. 215-247). Mahwah, NJ: Lawrence Erlbaum.Google Scholar
  44. Steglich, C., Snijders, T. A. B., and West, P. (2006). Applying SIENA: An illustrative analysis of the coevolution of adolescents’ friendship networks, taste in music, and alcohol consumption. Methodology, 6, 48–58. doi:10.1027/1614-1881.2.1.48 CrossRefGoogle Scholar
  45. Steglich, C. E. G., Snijders, T. A. B., and Pearson, M. A. (2010). Dynamic networks and behavior: Separating selection from influence. Sociological Methodology, 40, 329–393.CrossRefGoogle Scholar
  46. Stevenson, H. W., and Zusho, A. (2002). Adolescence in China and Japan: Adapting to a changing environment. In B. B. Brown, R. W. Larson and T.S. Saraswathi (Eds.). The world’s youth: Adolescence in eight regions of the world (pp. 141–170). Cambridge, UK: Cambridge University Press.Google Scholar
  47. Sun, W., Andreeva, B. A., Unger, J. B., Conti, D. V., Chou, C. P., Palmer, P. H., Sun, P., and Johnson, C. A. (2006). Age-related smoking progression among adolescents in China. Journal of Adolescent Health, 39, 686–693.CrossRefPubMedGoogle Scholar
  48. Veenstra, R., Dijkstra, J. K., Steglich, C., and Van Zalk, M. H. W. (2013). Network-behavior dynamics. Journal of Research on Adolescence, 23, 399–412. doi:10.1111/jora.12033 CrossRefGoogle Scholar
  49. Wen, X., et al. (2007). Modifiable family and school environmental factors associated with smoking status among adolescents in Guangzhou, China. Preventive Medicine, 45(2), 189–197. doi:10.1016/j.ypmed.2007.02.009 CrossRefPubMedGoogle Scholar
  50. West, P., and Michell, L. (1999). Smoking and peer influence. In P. West and L. Michell (Eds.), Handbook of pediatric and adolescent health psychology (pp. 179–202). Boston, MA: Allyn and Bacon.Google Scholar
  51. Zakriski, A. L., and Coie, J. D. (1996). A comparison of aggressive-rejected and nonaggressive-rejected children’s interpretations of self-directed and other-directed rejections. Child Development, 67, 1048–1070. doi:10.2307/1131879 CrossRefPubMedGoogle Scholar
  52. Zhang, L., Wang, W., Zhoa, Q., and Vartiainen, E. (2000). Prosocial predictors of smoking among secondary school students in Henan, China. Health Education Research: Theory and Practice, 15, 415–422. doi:10.1093/her/15.4.415 CrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media New York 2016

Authors and Affiliations

  • Ling Li
    • 1
  • Ting Lu
    • 2
  • Li Niu
    • 2
  • Yi Feng
    • 2
  • Shenghua Jin
    • 1
  • Doran C. French
    • 2
  1. 1.School of PsychologyBeijing Normal University and a Professor of Psychology at Hubei UniversityBeijingChina
  2. 2.Department of Human Development and Family StudiesPurdue UniversityWest LafayetteUSA

Personalised recommendations