The association between gender roles and smoking initiation among women and adolescent girls

Abstract

Smoking cigarettes is a leading global cause of preventable death and disease. Men historically smoke more than women, but the prevalence of smoking among women in low and middle-income countries is increasing at an alarming rate. Understanding the factors that influence smoking initiation among women and girls is needed to address the growing epidemic of women smokers and the looming impact on women’s health worldwide. We assume that smoking initiation is embedded in socio-culturally influenced gendered context and use a social cognitive model with a gendered lens as a framework for organizing and synthesizing the research. Guided by this framework, we identified gaps in the literature and make recommendations for future research in this review paper. The results suggest that psychological and environmental determinants are rooted in fluctuating cultural influences and values, but few research studies provide a gendered analysis or systematically examine these factors in the context of gender and culture. Sex/gender is a significant construct through which women and girls experience the psychological, environmental, and other influences on smoking initiation. Much more research is needed to understand the psychological and environmental influences as well as the intersection of gender roles and other social categories on female smoking initiation.

Introduction

Smoking cigarettes is a leading cause of preventable death and disease globally (WHO, 2015; US Department of Health and Human Services, 2014). Nonetheless, about 1 billion people in the world smoke tobacco, 80% of whom reside in low and middle-income countries (WHO, 2015) where men are historically much more likely to smoke than women. For example, in 2015 in Albania the prevalence of smoking was 10% among women and 40% among men; in Haiti, the prevalence was 2.5% among women and 22% among men (WHO, 2015). The prevalence of smoking among women in low and middle-income countries, however, is steadily increasing and the absolutely number of women who smoke worldwide is increasing at an accelerating pace (WHO, 2015). In 2001, women were estimated to comprise 20% of the world’s smoking population 2025 (Samet & Yoon, 2001); however, this milestone was achieved 6 years earlier, in 2018. We must have a better understanding of key factors that influence smoking initiation among women and girls to address the growing epidemic of women smokers and the looming impact on women’s health worldwide.

According to the tobacco epidemic model proposed by Lopez and colleagues (1994), global sex/gender differences in the prevalence of smoking are related to four non-linear stages associated with social and cultural context. In stage one, early in a country’s tobacco use trajectory, tobacco use increases among men and remains low among women. In stage two, smoking gains societal acceptance, men’s smoking reaches its peak and women’s smoking rapidly increases. During this stage, tobacco control efforts do not adequately contain tobacco use. In stage three, tobacco control efforts gain momentum, men’s smoking rates begin to decline while smoking among women increases then slowly starts to decline. In this stage, smoking becomes loses societal acceptance, especially among the well-educated. In stage four, smoking rates among men and women decrease until they become almost equal. In the model, evidence-based tobacco control efforts associated with declines in smoking prevalence rates include legislation, policies, and media campaigns (Lopez, Collishaw & Piha, 1994). Low and middle-income countries are often in stages one or two primarily due to a lack of effective tobacco control methods. The model emphasizes the need for a sex/gender-based focus on understanding and addressing tobacco use among women and girls because the systematic sex/gender variation of smoking behaviors are likely to have numerous biopsychosocial influences that need to be understood in the context of social and cultural influences.

Recognizing the need for sex/gender-based understandings of tobacco use, a growing body of research has gathered evidence for sex/gender-based biological, social, and psychological differences in smoking behaviors. Biologically, smoking cigarettes results in different patterns of brain stimulation and dopaminergic activation among women compared to men (Cosgrove, 2014; Beltz, Berenbaum & Wilson, 2015). Women’s menstrual cycles contribute to significant changes in smoking patterns as well as the ability to quit and stay quit (Saladin et al., 2015). Socio-cultural factors affect when, where and how much women smoke. For example, because smoking is not socially acceptable for women in South Korea, women become “closet” smokers, secretly smoking in their homes and denying it publicly (French et al., 2013). Also, light or intermittent smoking is an increasing trend of smoking and more common among females, young people and people from a higher socioeconomic background (Kotz, Fidler & West, 2012; Shiffman et al., 2012). Women also differ from men in the number of quit attempts they make, their relative quit success, and their responses to different cessation medications and behavioral interventions (Jarvis, Cohen, Delnevo, & Giovino, 2013; Smith et al., 2015). The health consequences of smoking also different. Women are significantly more prone than men to several tobacco-related diseases and health conditions including lung cancer and negative reproductive outcomes (US Department of Health and Human Services, 2014). Importantly, smoking among women has significant negative ramifications for the next generation. Women’s smoking during pregnancy has a substantial negative impact on fetuses and babies (US Department of Health and Human Services, 2014). Although women’s smoking is recognized as an important area of study and this body of work is growing, the findings still fall significantly short of providing substantive guidance in stemming the global epidemic of smoking among women (U.S. Department of Health and Human Services [USDHHS] 1980). There is a significant need for theoretically guided synthesis of research associated with smoking initiation among women and girls to identify gaps in knowledge and guide future research in the prevention of smoking among women and girls.

A gendered perspective theoretically grounded in the development of gender norms and roles would be particularly useful for developing a synthesis of research because gender roles are deeply rooted in social and cultural influences. These social and cultural influences begin from childhood within family environment (Carter, 2014). According to gender schema theory, proposed by Sandra Bem (1981), a society assigns specific tasks and roles for men and women. Children assimilate these tasks and roles via socialization and organize a gender-typed network of associations, called schema, to acquire and process information, guide their perception and psycho-social development across the life-course. According to sex role socialization theory, we start gender socialization within our family environment by learning and internalizing acceptable behaviors (Risman & Davis, 2012; Carter, 2014). What we learn is usually gendered and we automatically sex categorize everything in social relational contexts throughout life course, which results in gender becomes salient everywhere. That is stated as we do gender with our performances. Doing gender framework shifts the focus to individuals as the active agents who fulfill the expectations of feminine and masculine ‘natures’ with their practices and is the most common approach about gender in the literature (West & Zimmerman, 1987). From this viewpoint, researchers recently view gender as a social structure which is present in social relational contexts at individual, interactional and institutional levels (Risman, 2004). Gender is explicit at three dimensions: individuals develop gendered selves, interact men and women in social contexts based on cultural expectations. Also, regulations and structures of institutions are gender determined (Risman, 2004). As a result, gender is an ideological tool just as race and class. It is used as a background frame to conceptualize individual identities, interactions and institutions while enmeshed in other identities and activities in the foreground (West & Zimmerman, 1987; Risman, 2004; Carter, 2014).

The utility of using gender schemas and treating gender as a social structure when understanding tobacco use is perhaps best exemplified by tobacco industry marketing. Gender roles and perceptions are used to successfully promote tobacco products. Gender roles can be defined as gendered expressions or gender displays (West & Zimmerman, 1987). When smoking was only socially acceptable among men it was intentionally linked traditional masculine characteristics in western culture - masculinity, power, and independence. To promote smoking among women, smoking products were intentionally designed and linked with characteristics that were particularly meaningful to many women in western culture during that time, such as slimness, freedom, glamor, and attractiveness (Mackay & Amos, 2003; Anderson, Glantz & Ling, 2005). Tobacco companies targeted women by producing feminized cigarettes which are long, extra-slim, and menthol (Mackay & Amos, 2003). Advertisements also promote cigarettes via gendered psychosocial needs, unrelated to smoking using emotional messages to address women’s psychological needs such as “believe in yourself”, “spoil yourself”, and “ready to blow” with attractive, beautiful and feminine images (Anderson, Glantz & Ling, 2005).

Gender schema theory is likely to be helpful for considering how social/cultural differences in gender schema influence inter- and intra-societal variation in smoking initiation among women and girls. Social cognitive theory (SCT) has been used in health research to understand, explain, predict, and intervene with a variety of unhealthy behaviors, but is not been applied in a gendered manner to a synthesis of this research. Albert Bandura (2001) explains the process of learning in his theory (SCT) and claims that people are active agents who shape and are shaped by their environment. SCT can help to identify factors associated with understanding the initiation and maintenance of health-related behaviors (Wallston, 1992; Akers & Lee, 1996; Bidstrup et al., 2011). SCT posits that health behaviors are reciprocally influenced by personal factors (sex/gender, age, income, etc.), environmental factors (policies, media, family/friends smoking etc.), and behaviors (skills, practice, learning). Social cognitive models have been applied to other risky behaviors including sexual behavior, alcohol use, and smoking in the general population and is likely to be a strong conceptual framework for synthesizing the research associated with smoking initiation among women and girls (Akers and Lee, 1996; Bidstrup et al., 2011; Miller et al., 2015; Trujillo et al., 2015).

In this present paper, we reviewed studies on smoking initiation among women and girls using a gendered lens and used a social cognitive model as a framework for organizing the results. Guided by the framework, we identified gaps in the literature and made recommendations for future research.

Methods

Conceptual framework

We used a social cognitive model to guide our examination of the factors associated with the initiation of smoking among women and girls. SCT emphasizes reciprocal determinism, the interaction between individuals and all aspects of the environment, as well as the capacity for learning and adaptation. Elements in this social cognitive model are grouped into five categories, 1) psychological determinants of behavior, 2) observational learning, 3) environmental determinants of behavior, 4) self-regulation, and 5) moral disengagement (McAlister, Perry & Parcel, 2008). These components are described in detail below and reflected in Figure 1 which reflects multiple factors within the components that were hypothesized to affect smoking initiation among women and girls.

Figure 1.

Social Cognitive model of smoking initiation among women and girls

Psychological determinants are individual-level influences that affect the performance of a specific health behavior including outcome expectancies and self-efficacy. Outcome expectancies in composed of the perceived individual or social value of performing a specific behavior and include the anticipation of how one will feel about themselves if they perform a behavior as well as how one perceives that others will evaluate their behavior in addition to an individuals’ willingness to be influenced by these perceptions. Self-efficacy is individuals’ belief in their capacity to perform a specific behavior. These factors have been applied to understanding tobacco use in other populations (McAlister, Perry & Parcel, 2008).

Observational learning obtained through family, peers, and mass or remote communications, is another major component of SCT, and is particularly relevant to public health and tobacco use. Four processes influence observational learning (Bandura 1986): Attention, retention, production, and motivation. For smoking to become a learned behavior, individuals must attend to the behavior, retain a memory of the behavior, perceive an ability to perform the behavior, and be motivated to learn the behavior by having positive outcome expectancies.

Environmental determinants are environmental factors that influence the performance of a behavior. These include socio-demographic factors as well as inter-personal, familial, community, and societal factors. Some environmental factors are considered particularly powerful because without some environmental support (e.g., the availability of tobacco), women and girls cannot initiate tobacco use. The environment can provide rewards, punishments, or otherwise facilitate specific behaviors.

Self-regulation is the capacity to ensure short-term negative states in anticipation of long-term positive rewards. Sometimes included within psychological determinants, SCT views self-regulation as a product of motivation and the concrete skills for managing one’s own behavior. Moral disengagement is a process whereby individuals’ suspend moral standards for self-regulation.

We initiated the development of the model with several assumptions. Consistent with the concepts of intersectionality (Kelly, 2009; Hankivsky, 2012), socio-demographic variables such as sex/gender, age, education, etc. were identified as significant socially constructed categories that reflect asymmetrical life opportunities which in turn are associated with those structural inequalities. We also assumed that gender roles have a reciprocal relationship with tobacco policies, broad environmental influences (media and advertisement), immediate environmental influences (parent/sibling/peer smoking) and psychological influences. Gender roles established by society have an impact on the direction and power of these influences; while at the same time reinforce the strength of the gender roles. Lastly, psychological influences, which have reciprocal effects on gender roles, determine an individual’s access to cigarettes and the pathways to initiation and regular smoking behaviors.

We conducted a literature review and categorized findings into one or more of the components of the social cognitive model. For example, “psychological determinants of behavior” component of the theory involved articles from the “psychological influences” category in the model. “Observational learning” component has articles from “tobacco control policies”, “broad environmental influences” and “immediate environmental influences” categories of the model. “Significant social categories” were fitted into the “environmental determinants of behavior” component. Lastly, “self-regulation” and “moral disengagement” categories could be matched with articles from the “psychological influences” category; however, we did not find any relevant articles in the literature for these categories of the theory.

Search methodology

To synthesize the research, we conducted an extensive review of relevant gender-based scholarly works. This study was first initiated as a systematic review by using the PRISMA guidelines (Moher, Liberati, Tetzlaff, & Altman, 2009). The research question was, “How do gender roles influence smoking initiation among women in developed and developing countries?”. The inclusion criteria consisted of studies with samples of female participants, English-language studies, studies that have evaluated smoking initiation and studies that have assessed either gender roles or reasons for smoking among adolescent girls and women in general. The search terms were (tobacco OR cigarette OR nicotine OR smoking OR "nicotine dependence" OR "nicotine addiction") AND (female OR females OR woman OR women OR "gender role" OR "gender norms" OR "gender identity" OR girl OR girls OR feminine OR femininity) AND (uptake OR onset OR initiation) AND ("developed country" OR "developed countries" OR "developed nation" OR "developed nations" OR "industrialized nation" OR "industrialized nations" OR "industrialized country" OR "industrialized countries" OR "developing countries" OR "Third world country" OR "Third world countries" OR "Third world nation" OR "Third world nations". Seven electronic databases (Academic One File, CINAHL, Cochrane, Embase, PsychINFO, PubMed, Web of Science and Grey Literature) were searched using the key terms yielding n=293 potentially relevant publications. Two reviewers independently screened the titles and abstracts for inclusion. After this first step, n=240 articles were excluded because they didn’t meet the inclusion criteria, or they were unrelated to the research question; n=53 articles were identified for full-text review. The full text review yielded only thirteen relevant articles directly addressed the research question and perhaps another search strategy would be more productive. Research team members were assigned to specific components of the social cognitive model and searched the peer-reviewed published research articles for their assigned component using PsychINFO, PsychARTICLES, and PubMed. Each team member used the following key words: (tobacco OR cigarette OR nicotine OR smoking OR "nicotine dependence" OR "nicotine addiction") AND (female OR females OR woman OR women OR "gender role" OR "gender norms" OR "gender identity" OR girl OR girls OR feminine OR femininity) AND (uptake OR onset OR initiation) and also added a few more key words related to their own search area such as age or peer pressure. We concluded 57 relevant articles. Team members synthesized the findings independently and then as a group.

Results

Psychological Determinants of Smoking among Women and Girls

While there has been a significant amount of investigation into the psychological determinants of smoking in general, there has been little investigation into the psychological determinants of smoking initiation among women and girls and even less in different socio-cultural contexts. The few psychological constructs quantitatively assessed in relation to smoking initiation among women and girls are self-esteem, body image, eating pathology, and affective disturbances. In a group of adolescents (n=389) in British Columbia, 57% of whom were girls in grades 8 and 9, multivariate regression models were used to examine the extent to which self-esteem and gender and their interaction influenced the odds of smoking initiation. Self-esteem was assessed with the Rosenberg Self-Esteem Scale (RSES). For girls, the relationship between self-esteem and the initiation of smoking was significant. For each increase of one point on the RSES, the odds of initiating smoking among girls decreased by 9%, suggesting a potentially protective role of higher self-esteem (Richardson, Kwon & Ratner, 2013).

Increases in body dissatisfaction and eating pathology also appears to be associated with smoking initiation (Stice & Shaw, 2002) in addition to the belief that smoking helps control weight and conform to gendered stereotypes of female attractiveness (Stice & Shaw, 2003). Women’s self-esteem and stress levels can be negatively affected when they believe they don’t fit societal standards for beauty which also might result in smoking to control negative mood (Croghan et al., 2006). The tobacco industry has exploited these vulnerabilities and framed smoking as a way for women to feel beautiful, liberated, independent, empowered, and free (Hitchman & Fong, 2011).

The role of psychological factors also has been explored in two qualitative studies examining smoking among Australian Aboriginal women and educated Iranian women. The Australian Aboriginal women reported a mean age of smoking initiation of 15 years (range 12–21). Young girls reported that they initiated smoking to attain status, become or feel part of the group, and maintain their Aboriginal identity (Passey, Gale, & Sanson-Fisher, 2011). Iranian women reported initiation at a very young age and multiple reasons for initiation including immaturity, playfulness, and curiosity. Other factors included anticipating feelings of attractiveness, self-confidence, and independence. Some reported that smoking helped manage mood and provided a sense of tranquility and pleasure, but for others smoking was an expression of defiance of societal values and limitations (Baheiraei et al., 2016).

Observational Learning and Smoking Initiation among Women and Girls

Observational learning via parent, peers, and siblings has a powerful effect on adolescents’ smoking initiation (Akers and Lee, 1996; Bidstrup et al., 2011; Miller et al., 2015; Trujillo et al., 2015). Parental, sibling and peer smoking are significant risk factors of smoking initiation among adolescents (Peters et al., 2005; Scragg & Glover, 2007; Ravishankar & Nagarajappa, 2009; Harakeh & Vollebergh, 2012; Chassin, Presson & Sherman, 1995; West, Sweeting & Ecob, 1999). Although there are some gender differences in smoking initiation among adolescents, gender is not a consistent predictor of the smoking initiation (Villanti, Boulay & Juon, 2011). The prevalence and the age of smoking initiation among girls and boys varies across the world. Although boys are still more likely to experiment or initiate tobacco in developing countries, the initiation rate has been increasing among females since the 1940s and is now higher compared to males in some developed countries such as the US (Anderson & Burns, 2000; Ravishankar & Nagarajappa, 2009). These statistics might reflect changes in the cultural differences and gender norms towards smoking within the society across different countries, but much of the variance attributed to observational learning from parents, sibling and peers remains unexplained.

Some studies suggested modeling from parents to be the strongest predictor of smoking initiation for both boys and girls (Hoving, Reubsaet & de Vries, 2007); however, maternal and paternal smoking may have different effects on girls’ and boys’ smoking behaviors, although the findings are inconsistent and are likely to depend on socio-cultural context (Oygard et al., 1995; Scragg & Glover, 2007; Harakeh & Vollebergh, 2012). Maternal smoking was found to have a greater influence on their children’s smoking compared to paternal smoking (Oygard et al., 1995; Scragg & Glover, 2007; Harakeh & Vollebergh, 2012). Parental approval of smoking has been found to have a mediating effect of parental smoking for girls only and parental smoking was a stronger predictor of females’ smoking uptake (Flay et al., 1994; Oygard et al., 1995). Nonetheless, in some studies sibling smoking was found to have a stronger impact than parental smoking and a lesser impact than friends smoking on adolescents’ smoking uptake for both boys and girls during early adolescence (Oygard et al., 1995; West, Sweeting & Ecob, 1999).

Peer pressure has also been suggested as the main reason for smoking uptake by numerous studies (Peters et al., 2005; Ravishankar & Nagarajappa, 2009). The relationship between smoking initiation among adolescents and friends’ or peer smoking has been found to be dynamic and complex (Peters et al., 2005; Hall & Valente, 2007; Harakeh & Vollebergh, 2012). Friends are generally present during adolescents’ first experience with cigarette smoking (Peters et al., 2005) and peers have both immediate and future influence on adolescent smoking behaviors (Hall & Valente, 2007). The reciprocal nature of the relationship is also important. The selection of smokers as friends predicts future smoking behavior and susceptibility to smoke of adolescents (Hall & Valente, 2007). Moreover, passive peer influence (i.e., modeling) appears to have a greater influence on adolescents’ smoking behavior than active peer influence (i.e., peer pressure). Adolescents intentionally imitate friends and respected others without pressure (Harakeh & Vollebergh, 2012). Also, friends’ smoking directly affected adolescents’ smoking initiation and indirectly affected the maintenance of adolescents’ smoking behavior via the mediating effects of perceived friends’ approval of smoking, refusal self-efficacy, and negative outcome expectations and intentions (Flay et al., 1994). Developmental stage also appears to be an important consideration. Several studies suggest that peer smoking is a more salient predictor of the smoking initiation during early adolescence (Chassin, Presson & Sherman, 1995; West, Sweeting & Ecob, 1999; Villanti, Boulay & Juon, 2011). Although the literature provides a valuable information regarding the influence of peer pressure on adolescents’ smoking behaviors, related studies have not sought for gender differences. Therefore, we don’t know if girls and boys are differently influenced by their peers for tobacco use and in what extent.

Overall, girls are influenced by parental, sibling and friends smoking behaviors and the effects of these influences differ by age of exposure. The influence of parental smoking is more effective during the early years of adolescence and in appears to be temporary (Flay et al., 1994). Peer smoking has the greatest and yet the most complex influence on adolescents. Although its influence decreases by age, it still stands as the only significant and the strongest predictor of smoking initiation from early adolescence until adulthood (Flay et al., 1994; West, Sweeting & Ecob, 1999; Villanti, Boulay & Juon, 2011). Adolescents are directly influenced by peer and sibling smoking via imitating them, whereas parents’ smoking affects adolescents’ attitudes and beliefs toward smoking.

Environmental Determinants of Smoking among Women and Girls

Some environmental factors are associated with smoking initiation, and while a few have been shown to have greater influence on females, other factors have not been sufficiently studied or have not shown differences in terms of gender. Furthermore, although there are themes that cut across cultures and regions, there are also noticeable differences in factors associated with female smoking initiation between developed and developing countries and among different cultures. Age is perhaps the only consistent factor across countries and cultures, with women tending to start smoking later in adolescence and adulthood than men across diverse countries and socio-cultural contexts like US, Turkey, and Romania (Kaleta, Usidame, Dziankowska-Zaborszczyk, & Makowiec-Dąbrowska, 2015; Kara, Dikici, & Yel, 2010; Thompson, Tebes, & McKee, 2015). This trend is particularly problematic because later smoking initiation is also associated with more significant physical health risks for women versus men (Thompson et al., 2015), however the reasons for this are unclear.

One potential reason for later initiation in females could be an increase in individual income, which has been found to be associated with increased smoking in 15–24 year olds (Blakely, van der Deen, Woodward, Kawachi, & Carter, 2014), however there is little available data on individual income differences and workforce participation by gender in this age group. Furthermore, increases in household income were not associated with similar increases in smoking initiation, regardless of gender, and it was further found that an increase in household income that moved individuals across the poverty threshold was actually found to be related to decreases in smoking regardless of gender (Yoon et al., 2007), suggesting that improved socioeconomic status (SES) may have a positive impact on smoking behavior. However, the impact of improved SES is also not a consistent predictor of smoking initiation, with others finding increases in SES to be associated with greater likelihood of smoking (Kaleta et al., 2015), as well as finding that youth from countries with greater income inequality have greater likelihood of smoking and that increased GDP of a country increases the odds of being a current smoker in that country (Li & Guindon, 2013), suggesting that the relationship between SES and smoking initiation is complex.

A number of individual factors have been identified as potential moderators of smoking initiation, including education, employment status, marital status, sexual orientation, and religiosity. While in some cases there is little research examining differences in the influence of these factors on smoking intiation by gender, there are some identifiable trends. Specifically, unmarried, and particularly divorced individuals were found to be more likely to initiate smoking and exhibited higher nicotine dependence (Abiola et al., 2016; Lindström, 2010; Pennanen et al., 2014), while individuals with lower levels of education and lower parental education were associated with greater exposure to smoking, greater likelihood of smoking initiation, greater regular smoking and lower cessation rate (Federico, Costa, & Kunst, 2007; Morin, Rodriguez, Fallu, Maïano, & Janosz, 2012; Yáñez, Leiva, Estela, & Čukić, 2017; Zhuang, Gamst, Cummins, Wolfson, & Zhu, 2015). Unemployment, insecure employment, and recent involuntary job loss were also found to be significant risk factors for smoking initiation, decreased quitting, and smoking relapse after quitting regadless of gender (Golden & Perreira, 2015; Jung, Oh, Huh, & Kawachi, 2013; Ohlander, Vikström, Lindström, & Sundquist, 2006). In terms of sexual orientation, research has found that homosexual and bisexual youths were at greater risk for smoking initiation, and this was especially true for bisexual girls (Easton, Jackson, Mowery, Comeau, & Sell, 2008). This relationship is believed to be the result of greater daily stress associated with homophobia, discrimination, etc. based on stress theory, which argues that sexual minorities smoke to cope with difficulties arising from stigmatized sexual orientation (Corliss et al., 2014). The common thread among these individual factors seems to be stress, which has been previously identified as an important risk factor in smoking initiation, especially among females, who are also more likely to be diagnosed with depression and to smoke to manage negative mood (Columbus, 2001; Iakunchykova et al., 2015; “Women and smoking: a report of the Surgeon General. Executive summary.,” 2002). Finally, higher religiosity is associated with lower rates of smoking initiation, especially among females; high levels of religiosity also appear to moderate genetic factors associated with smoking initiation, however the rationale for this relationship is unclear (Kendler et al., 1999; Timberlake et al., 2006; Whooley, Boyd, Gardin, & Williams, 2002). These relations are speculatively explained by the social benefits of being part of the religious community helping to cope with stress in addition to constraining smoking behaviors.

The primary environmental factors associated with smoking intiation tend to vary significantly depending on country and culture, however adolescent girls tend to more susceptible to social influences on smoking initiation than boys (Hu, Flay, Hedeker, Siddiqui, & Day, 1995). Smoking initiation among females in cultures with strong collectivist familial social structures, such as Turkey, Iran, and Colombia, appears associated with familial factors such as mother’s education level, family smoking rituals, and the quality of parental relationships (Afanador, Radi, Pinto, Pinzón, & Carreño, 2014; Baheiraei et al., 2015; Kara et al., 2010). In cultures that support greater individualism, often located within higher per capita income countries, such as the US and Netherlands, smoking initiation among females appears associated with factors like self-image, weight control, delinquency, attention problems, a sense of rebellion against unconventional values, a feeling of independence, and social acceptance (de Meer, Crone, & Reijneveld, 2010; “Women and smoking: a report of the Surgeon General. Executive summary.,” 2002). An assessment of Hispanic/Latino adolescents (aged 10–18) in the US found that risk of smoking initiation was two times higher for women born in the US than those born outside the US, and even among non-US born women, smoking initiation risk increased with duration of residence in the US (Parrinello et al., 2015).

Although the socio-cultural context varies, these findings highlight the significant effects of societal and cultural expectations on smoking initiation. These expectations also vary within countries and across race and ethnicity. For example, peer smoking and peer group identification are consistently strong predictors of smoking initiation among white adolescents, but this relationship is found less consistently among Hispanics and far less consistently among African Americans (Mermelstein, 1999). Among white adolescent girls, common beliefs about smoking include helping to control weight and enhancing one’s image of being independent or sophisticated, while common beliefs among African American adolescent girls include fear that smoking might negatively affect employment opportunities and advancement, and that others will inappropriately attribute other negative behaviors to African Americans who smoke (Mermelstein, 1999). Unsurprisingly, identification with mainstream white society is linked with greater regular smoking among Hispanic women (Marin, Perez-Stable, & Marin, 1989). These finding highlight the potential for peer group influences to overcome preventative family and culture influences.

Tobacco control policies also affect smoking initiation of women and girls. In general, smoke-free workplaces and especially smoke-free home policies prevent smoking initiation among adolescents (Farkas et al., 2000; Muilenburg et al., 2009; Gorini et al., 2016). However, there are no known sex/gender differences on the impact of tobacco control policies primarily because the topic has not as yet been explored. Nonetheless, women tend to smoke at home in socio-cultural contexts were smoking is unacceptable for women (French et al., 2013), suggesting that a smoke-free home policies are not in place in these contexts and might have an impact on women’s smoking if implemented.

Self-Regulation

While sex/gender differences in smoking initiation has not been explored within the large body of literature focused on tobacco use and self-regulation, the experience of being marginalized is linked to a greater focus on attaining pleasure in the present than developing goals for the future (Passey, Gale & Sanson-Fisher, 2011) and might have an effect on female smoking in many cultures that marginalize women’s human rights.

Discussion

This project used a social cognitive model as a framework and a gendered perspective for organizing and synthesizing the research on smoking initiation among women and girls. We reviewed relevant studies in the literature, used the model to categorize findings, identified significant gaps in the literature, and make recommendations for future research.

Sex/gender is a significant construct through which women and girls experience the psychological, environmental, and other influences on smoking initiation. Psychological influences that specifically impact women’s smoking include self-esteem, body dissatisfaction, eating pathology, mood management, attractiveness, and feelings of independence, empowerment, and freedom (Stice and Shaw 2002; Stice and Shaw 2003; Richardson, Kwon & Ratner, 2013). Observational learning is a mechanism thought which social factors in the environment impact smoking initiation. The most studied influences include parental, sibling, and peer relationships. Each of these groups has a different impact on smoking at different stages in of adolescence (Peters et al., 2005; Scragg & Glover, 2007; Ravishankar & Nagarajappa, 2009; Harakeh & Vollebergh, 2012). In addition, the effect of parental smoking appears to vary based on the parent’s sex/gender which is likely to be related to traditional roles of mothers and fathers. Mothers are expected to be the first resource for children and their behaviors and as such might be a more powerful model for female children’s behaviors. Moreover, smoking has been accepted as a male behavior for many years. Adolescents who assimilated this belief are likely to perceive that smoking belongs to fathers and maternal smoking breaks this norm which might make the behavior more salient for girls. Girls are more affected by parental smoking than boys (Flay et al., 1994; Oygard et al., 1995). We speculate that this might be associated with more girls being raised to be family-oriented than boys. Virtually no research has examined the influences of these relations on the initiation of smoking among adult women which is important because females tend to initiate smoking at later ages than males across the world. Future research needs to examine whether the internalization of women’s gender roles within a society might mediate these findings.

The relation between SES and smoking initiation among women varies considerably across the world (Kaleta, Usidame, Dziankowska-Zaborszczyk, & Makowiec-Dąbrowska, 2015; Kara, Dikici, & Yel, 2010; Thompson, Tebes, & McKee, 2015). Other social categories that increase the likelihood of smoking initiation include being unmarried or divorced, unemployed, with insecure employment, having a recent involuntary job loss, homosexuality, and bisexuality (Abiola et al., 2016; Lindström, 2010; Pennanen et al., 2014; Golden & Perreira, 2015; Jung, Oh, Huh, & Kawachi, 2013; Ohlander, Vikström, Lindström, & Sundquist, 2006); however, these factors are linked to an increased stress levels and women do not appear to differ from men with respect to these influences on smoking (Columbus, 2001; Iakunchykova et al., 2015; “Women and smoking: a report of the Surgeon General. Executive summary.,” 2002). Religiosity appears to be a protective factor against smoking initiation among females (Kendler et al., 1999; Timberlake et al., 2006; Whooley, Boyd, Gardin, & Williams, 2002).

The primary limitation in this initial examination is simply a lack gendered perspectives in the research in smoking initiation among women and girls, especially with a systematic focus on gender norms and cross-cultural differences. Because there are so few studies that directly assess sex/gender differences and incorporate gender roles, our conclusions are preliminary and are put forth as starting point for future research. Because the systematic review returned too few studies and we conducted instead a broad scoping review, we included some studies for which gender was not the primary focus of the study.

In conclusion, smoking initiation among women and girls is increasing at an alarming rate and we need a better understanding of key factors that influence this situation to address a growing worldwide epidemic of women smokers and the looming impact on women’s health. Much more research is needed to understand the psychological and environmental influences as well as the intersection of gender roles and other social categories on female smoking initiation.

Article name	Author	Year	Sample	Outcomes
Age of initiation, Determinants and Prevalence of Cigarette Smoking among Teenagers in Mushin Local Government Area of Lagos State, Nigeria	Abiola, et al.	2016	475 teenagers	Mean age of smoking uptake is 12. Age, gender, marital status, educational background, friends and media. Unmarried and divorces individuals were more likely to be smokers and have a higher level of nicotine dependence.
Sociocultural Determinants of Tobacco Smoking Initiation among University Students in Bucaramanga, Colombia, 2012	Afanador et al.	2014	College students (Case group: 167 and control group: 314)	Environmental factors are determinants of smoking uptake: being away from hometown, girlfriend/boyfriend smoking, having a bad relationship with father and having a mother who smoke or drink alcohol.
A Longitudinal Test of Social Learning Theory: Adolescent Smoking	Akers & Lee	1996	454 students in grades 7 through 12 (5-year longitudinal data from the Iowa study)	Social learning variables have a stable effect on smoking behavior. Social reinforcements for smoking initiation are peer and parental influences. Girls are influenced by their boyfriends’ smoking more than boys are influenced by their girlfriends’ smoking.
Patterns of adolescent smoking initiation rates by ethnicity and sex	Anderson & Burns	2000	Hispanic, non-Hispanic, white and African American adolescents aged 12-17 years for each calendar between 1940-1992	Initiation rates among males in all ethnic groups have declined since 1945; however, it has increased since 1983. Initiation rates among females have increased since 1940 and cought up to males’ initiation rates by the mid-1970s.
The Role of Family on Hookah Smoking Initiation in Women: A Qualitative Study	Baheiraei et al.	2015	36 Iranian women	Presence of hookah in home environment and girls’ involvement in the preparation of hookah predicts the initiation of hookah among women.
Experiences of cigarette smoking among Iranian educated women: A qualitative study.	Baheiraei et al.	2016	14 educated Iranian female smokers	Personal factors (imitation, show-off, independence, curiosity, personal interest and desire, improved mood and social defiance), family factors (family smoking, family conflicts and family limitations) and social factors (work and school environment, gender equality symbols, peer pressure and acceptance) are the main influential factors of smoking behavior.
Critical discussion of social–cognitive factors in smoking initiation among adolescents	Bidstrup et al.	2011	12 Danish adolescents in grades 7 and 8	Peer pressure, friends’ smoking and parental smoking were predictors of smoking uptake among adolescents.
Do changes in income, deprivation, labour force status and family status influence smoking behaviour over the short run? Panel study of 15,000 adults	Blakely et al.	2014	15000 adults in two observations over three waves (two years apart)	Personal income increase was associated with increase in smoking for 15-24 years olds. Worsening social circumstances were linked to increased odds of smoking.
Social-psychological antecedents and consequences of adolescent tobacco use	Chassin et al.	1995	The follow-up study consisted of 8521 adolescents and young adults between 15-24 years of ages.	The initiation of smoking was better predicted by social environments whereas transition to regular smoking was better predicted by the subjects’ attitudes and beliefs about smoking. Peer and parental influences were significant for adolescents’ smoking.
Advances in psychology research	Columbus	2001
Sexual orientation disparities in adolescent cigarette smoking: intersections with race/ethnicity, gender, and age	Corliss et al.	2014	64397 students in grades 9-12	Sexual minority adolescents were more likely to be smokers compared to their heterosexual peers. Bisexual adolescent girls, Black gays and lesbians and adolescent girls were at the most risk.
Is smoking related to body image satisfaction, stress, and self-esteem in young adults?	Croghan et al.	2006	2057 students between 18-24 years of age.	Higher perceived stress and low self-esteem were significant predictors of smoking behavior regardless of gender. Female smokers had higher level of body dissatisfaction compared to non-smoker females and male smokers.
Gender differences in the association between pre-adolescent smoking initiation and emotional or behavioral problems	de Meer et al.	2010	960 children between the ages of 7-13.	Smoking behavior was associated with delinquency, attention problems, and thought problems for girls. Boys’ smoking was not associated with problem behaviors.
Adolescent Same-Sex and Both-Sex Romantic Attractions and Relationships: Implications for Smoking	Easton et al.	2008	Data of 12603 adolescents (6203 boys and 6400 girls) from the National Longitudinal Study of Adolescent Health	LGB youth were at higher risk of smoking.
Association between household and workplace smoking restrictions and adolescent smoking	Farkas et al.	2000	Data of 17185 adolescents at the ages of 15-17 from national population-based surveys conducted in 1992-1993 and 1995-1996.	Smoke-free homes and smoke-free workplaces were protective factors against adolescent smoking behaviors where smoke-free homes had a greater influence compared to smoke-free workplaces.
Educational inequalities in initiation, cessation, and prevalence of smoking among 3 Italian birth cohorts	Federico et al.	2007	Data from 1999-2000 national survey (28958 men and 29769 women born between 1940-1969)	Socio-economic and educational inequalities in the lifetime increased the risk of being a smoker.
Differential influence of parental smoking and friends' smoking on adolescent initiation and escalation of smoking	Flay et al.	1994	6695 seventh grade students (data collected in 1986)	Peer smoking had indirect and direct influences on adolescent smoking initiation whereas parental smoking had only indirect effects among adolescents. Parental approval of smoking affected only girls and parental smoking had stronger influence for females.
Cross-national gender differences in the socioeconomic factors associated with smoking in Australia, the United States of America and South Korea	French et al.	2013	Cross-sectional data conducted from national surveys in 2006	Males had a greater smoking rates than females in all countries, but S Korea had the greatest gender difference. Low educational level increased the odds of smoking except Korean women. Gender empowerment had an inverse relationship with gender differences in smoking.
Losing jobs and lighting up: Employment experiences and smoking in the Great Recession	Golden & Perreira	2015	Data of 13571 people from a longitudinal national survey between 2001-2011	Involuntary job loss, unemployment and unsecure job status promoted smoking uptake.
Smoke-free homes and youth smoking behavior in Italy: findings from the SIDRIAT longitudinal study	Gorini et al.	2016	778 children aged 6–7 years and 985 adolescents aged 13–14 in 2002	Smoke-free home policies had a protective influence against smoking initiation among children and adolescents
Adolescent smoking networks: the effects of influence and selection on future smoking	Hall & Valente	2007	880 participants in grades between 6 and 7	The influence of peer smoking is a process for smoking uptake. Selecting smokers as friends had a direct effect for smoking behavior and selecting susceptible smokers as friends had a future effect to initiate smoking.
The impact of active and passive peer influence on young adult smoking: an experimental study	Harakeh & Vollebergh	2012	68 daily smokers aged 16-24	Peer influence had a significant effect on adolescent for smoking behavior via modeling whereas peer pressure did not.
Gender empowerment and female-to-male smoking prevalence ratios	Hitchman & Fong	2011	Data of male and female smokers older than 15 years of age from the WHO report in 2008	Women’s empowerment (social changes, women’s economic and political participation, and women’s rising economic resources) increased women’s smoking ratio and narrowed the gap between men’s and women’s smoking prevalences.
Predictors of smoking stage transitions for adolescent boys and girls	Hoving et al.	2007	4055 non-smokers (7th grade students) at baseline from six European countries	Girls were more likely to start smoking. Parental smoking is the strongest predictor of smoking uptake. Perceived pros of smoking for girls is related to weight control. Smoking uptake for girls was significantly predicted from perceived pressure to smoke from friends
The Influences of Friends’ and Parental Smoking on Adolescent Smoking Behavior: The Effects of Time and Prior Smoking	Hu et al.	1995	6695 adolescents in 7-9 grades. Data was collected in 4 times between 1986-1988	Parental and peer smoking were predictors of adolescent smoking. The influence of peer smoking was stronger and increased over time where the influence of parental smoking was stable.
The impact of early life stress on risk of tobacco smoking initiation by adolescents	Iakunchykova et al.	2015	1020 pregnant women from Ukraine who were pregnant and followed-up with their kids until they were 16 years old	Early life stressors (physical or sexual abuse, being separated from someone, maternal alcohol use, and death of a family member etc.) were predictors for early smoking initiation for adolescents.
The effects of employment conditions on smoking status and smoking intensity: the analysis of Korean labor camp; income panel	Jung et al.	2013	1373 Korean males aged 20-59	Insecure employment was a stronger predictor of smoking uptake compared to unemployment
Socioeconomic Disparities in Age of Initiation and Ever Tobacco Smoking: Findings from Romania	Kaleta et al.	2015	Data from Global Adult Tobacco Survey in Romania	Unemployment, age and having a high asset were predictors for smoking uptake.
Smoking Initiation in Primary School Students in Southeast of Turkey: the Roles of Sociodemographic Factors, Gender and Parental Characteristics	Kara et al.	2010	1124 adolescents	Smoking became more prevalent in primary schools where boys more tended to smoke. Peers and relatives influence adolescent smoking uptake.
A population-based twin study in women of smoking initiation and nicotine dependence	Kendler et al.	1999	Caucasian female same-sex twins (N = 2163)	Both environmental and genetic factors (lower educational level and religiosity, high levels of neuroticism and extroversion, and a history of psychological disorders) affected smoking initiation.
Income, income inequality and youth smoking in low- and middle-income countries	Li & Guindon	2013	Data from Global Youth Tobacco Surveys. 169,283 adolescents aged 13-15 from low and middle income countries.	Level of income and smoking initiation was positively related among youth. Unequal distribution of income promoted smoking uptake.
Social capital, economic conditions, marital status and daily smoking: a population-based study	Lindström	2010	7,757 participants between 18–80 years of age	Never married and divorced people were more likely to start smoking.
Cigarette smoking among San Francisco Hispanics: the role of acculturation and gender	Marin et al.	1989	1669 Hispanics aged 15-64.	More accultured Hispanics tended to smoke more.
Ethnicity, gender and risk factors for smoking initiation: an overview	Mermelstein	1999	This was a review paper.	White females were more susceptible to the negative influence of their friends. African American parents sent stronger anti-smoking messages to their kids compared to White parents. African American females perceived more negative outcomes of smoking.
Using social cognitive theory to predict intention to smoke in middle school students	Miller et al.	2015	163 middle school students	Intention for not smoking among students were influenced by several social cognitive factors: emotional coping for not smoking, environmental support for not smoking, expectations for not smoking and self-control for not smoking.
Academic achievement and smoking initiation in adolescence: a general growth mixture analysis	Morin et al.	2012	741 nonsmokers interviewed from grade 7th to 10th.	Low or unstable GPA was associated with higher risk of smoking uptake. However, girls had higher GPA over time and were at higher risk of smoking.
The home smoking environment: influence on behaviors and attitudes in a racially diverse adolescent population	Muilenberg et al.	2009	4,296 students in a southern urban area	Smoking free home policies affect smoking behaviors, beliefs, susceptibility to smoking and motivation to quit.
Neighbourhood non-employment and daily smoking: a population-based study of women and men in Sweden	Ohlander et al.	2006	31,164 participants aged 25–64	There was a positive relationship between neighborhood unemployment and daily smoking. Young women in the lowest occupational groups were at higher risk of smoking.
Parental and peer influences on smoking among young adults: ten-year follow-up of the Oslo youth study participants	Oygard et al.	1995	827 students aged 11-14 and their parents in 1979	Friends and sibling smoking were associated with smoking uptake. Mother’s smoking was the strongest predictor of smoking.
Risk of Cigarette Smoking Initiation During Adolescence Among US-Born and Non-US-Born Hispanics/Latinos: The Hispanic Community Health Study/Study of Latinos	Parrinello et al.	2015	16 415 Hispanic/Latino adults	Women born and residing in the US were at higher risk of smoking initiation than women living outside of the US did.
"It's almost expected": rural Australian Aboriginal women's reflections on smoking initiation and maintenance: a qualitative study	Passey et al.	2011	3 focus groups with 14 Aboriginal women and service providers and 22 individual interviews with Aboriginal women	Girls smoke to attain status and to maintain their belonging to Aboriginal identity. Socio-economic disadvantage, housing shortage and unemployment promote smoking behaviors.
Smoking, nicotine dependence and nicotine intake by socio-economic status and marital status	Pennanen et al.	2014	1746 ever smokers	Lower levels of education, SES and occupation were associated with higher risk of smoking uptake.
Beliefs and social norms about smoking onset and addictions among urban adolescent cigarette smokers	Peters et al.	2005	52 current smoker high school students	Friends and sibling smoking impacted adolescents’ smoking initiation.
Factors attributing to initiation of tobacco use in adolescent students of Moradabad, (UP) India	Ravishankar & Nagarajappa	2009	590 adolescents	Curiosity, peer pressure, parental smoking and especially parents’ smoking at home were associated with experimental and daily smoking
Self-esteem and the initiation of substance use among adolescents	Richardson et al.	2013	Data of 1267 adolescents in grades 8-9 from the British Columbia Adolescent Substance Use Survey in 2010-2011	Self-esteem was negatively associated with the initiation of tobacco use which the relationship was stronger for girls.
Parental and adolescent smoking: does the association vary with gender and ethnicity?	Scragg & Glover	2007	Data of 91,219 students from 2002-2004 surveys	Parental smoking was a risk factor for adolescents’ smoking initiation where maternal smoking had a greater effect.
Role of body dissatisfaction in the onset and maintenance of eating pathology: a synthesis of research findings	Stice & Shaw	2002	This was a review paper	Internalization of thin-ideal body image led to body dissatisfaction which in turn promotes eating pathology and smoking initiation
Prospective relations of body image, eating, and affective disturbances to smoking onset in adolescent girls: how Virginia slims	Stice & Shaw	2003	496 adolescent girls	Body dissatisfaction and eating pathology predicted the onset of smoking
Gender differences in age of smoking initiation and its association with health	Thompson et al.	2015	16,985 adults from 2001-2002 survey	Women surpassed men in smoking initiation in late adolescence and adulthood. Late initiation was linked to physical health problems for women more than men. Early initiation was linked to major depressive disorder and anxiety disorder for women.
The moderating effects of religiosity on the genetic and environmental determinants of smoking initiation	Timberlake et al.	2006	237 monozygotic twin pairs, 315 dizygotic twin pairs, 779 full-sibling pairs, and 233 half-sibling pairs in young adults (from National Longitudinal Study of Adolescent Health data)	Higher levels of self-rated religiousness was associated with lower prevalence of smoking initiation.
How adolescents learn about risk perception and behavior in regards to alcohol use in light of social learning theory: A qualitative study in Bogotá, Colombia	Trujillo et al.	2015	160 students in grades 7-8	Social learning theory determinants (modeling from parents and conveying media messages) impacted adolescents’ health risk behaviors.
Peer, parent and media influences on adolescent smoking by developmental stage	Villanti et al.	2011	Data from the National Youth Tobacco Survey (2004)	Peer smoking, smoking at home, exposure to tobacco industry and media was associated with smoking in early and middle adolescence.
Family and friends' influences on the uptake of regular smoking from mid-adolescence to early adulthood	West et al.	1999	1009 15-year-olds interviewed and followed-up at 16, 18, 21 and 23 years of age (1987-1996)	Peer smoking had a greater impact on smoking uptake than family smoking. Peer influence lessened over time but it maintained significant.
Religious involvement and cigarette smoking in young adults: the CARDIA study (Coronary Artery Risk Development in Young Adults) study	Whooley et al.	2002	4569 adults between the ages of 20-32	Attending religious services was a protective factor for smoking especially for young females
The associations of personality traits and parental education with smoking behavior among adolescents	Yáñez et al.	2017	842 adolescents between 14-15 years of age	Adolescents with higher levels of extraversion and neuroticism and parents with lower educational background had higher rates of smoking. Higher level of conscientiousness was protective against smoking uptake.
Delay discounting predicts postpartum relapse to cigarette smoking among pregnant women	Yoon et al.	2007	48 pregnant women	Delay discounting was significantly related to smoking status at 24 weeks postpartum.
Comparison of smoking cessation between education groups: findings from 2 US National Surveys over 2 decades	Zhuang et al.	2015	Data from the National Health Interview Survey (1991–2010) and the Tobacco Use Supplement to the Current Population Survey (1992–2011).	Lower educational levels was associated with lower quit attempt and lower success rate at quitting.

Biography

Nurbanu Ozbay, received her BA degree from Hacettepe University, Turkey (2013) and her MA from the General Psychology Program at City College New York (2017). She worked as a research assistant on Dr. Lesia Ruglass’ (CCNY) and Dr. James Root’s (MSKCC) pilot research project examining increased cue reactivity to smoking-related cues among Black Americans. Her research interests include female psychology, trauma, tobacco disparities and culture.

Christine E. Sheffer, PhD, is a Clinical Psychologist and an Associate Professor of Oncology at Roswell Park Comprehensive Cancer Center. She is also the President of the Council for Tobacco Treatment Training Programs and the Director of Research and Evaluation for Roswell Park Cessation Services, which provides tobacco treatment Quitline services for the state of New York, New Jersey, and several health care insurance providers. As a clinician, researcher, and educator, her goals include contributing to incremental increases in the efficacy and uptake of high-quality evidence-based treatments for tobacco dependence, enabling individuals to obtain the help they need to achieve and maintain good health. Her research examines bio-behavioral mechanisms that initiate and sustain tobacco dependence, and methods to improve the efficacy and uptake of empirically-supported treatments for tobacco use, particularly among groups that experience tobacco-related health disparities. Dr. Sheffer previously served as faculty in the City University of New York Medical School and the University of Arkansas for Medical Sciences.

Alina Shevorykin, is a PhD student in Mental Health Counseling at Pace University and she completed the Master’s degrees in Mental Health Counseling and General Psychology at The City College of New York. Her undergraduate study included Economics and Linguistics. Alina’s research focuses primarily on the links between trauma, resiliency and substance use. Currently, she is studying interventions aimed at tobacco cessation, understanding how traumatic experiences can influence substance use and treatment outcomes, as well as electrophysiological and cognitive mechanisms underlining tobacco and cannabis addictions.

Philip Smith, PhD, is an Assistant Professor with the Miami University Department of Kinesiology and Health. His research focuses on behavioral mechanisms of addiction and tobacco use disparities from an intersectional framework.