Smoking Susceptibility and Anti-Smoking Awareness in Adolescents and Young Adults of Bangladesh
Article information
Abstract
Objective
The present study addressed the shortcomings of strictly defined criterion and survey-based approaches of previous smoking susceptibility measures and accordingly, developed and validated two scales, Smoking Susceptibility Measure (SSM) and Anti-Smoking Awareness Scale (ASAS).
Methods
Firstly, the generation of SSM and ASAS items followed an extensive literature review, expert opinions and agreement, resulting in the retention of eight items for SSM and seven items for ASAS to administer them on a large sample (n=312). Average inter-item correlations, corrected item-total correlations, and internal consistency reliabilities of the measures fall within the recommended ranges. The data were found suitable to factorize the sample through exploratory factor analysis. To determine the structural validity of the measures, confirmatory factor analysis (CFA) was done, and the data had an adequate model fit for unifactorial solution. Multi-group CFA revealed that both measures can be applied in the same way across age and sex of the participants.
Results
An inverse association of school connectedness with smoking susceptibility and positive association with anti-smoking awareness reflected the convergent validity of the measures. Hierarchical regression analysis showed that smoking susceptivity was negatively predicted by anti-smoking awareness and school connectedness whereas, positively predicted by self-esteem of the participants.
Conclusion
The SSM and ASAS were found to be psychometrically sound tools to objectively measure never-smoking youths’ smoking susceptibility and anti-smoking awareness, targeting effective intervention strategies to prevent adolescents and young adults from being a regular smoker.
INTRODUCTION
With over one billion smokers globally, tobacco use is a growing concern and a significant cause of preventable mortality. Therefore, urgent action is imperative to address this epidemic and prevent tobacco-related death in the 21st century [1,2]. As a dominant form of tobacco usage worldwide, the prevalence of cigarette smoking is at the top of the hierarchy; furthermore, 90% of smokers are adolescents and experience smoking between the ages of 10 and 20 years [3,4]. In Bangladesh, smoking is widely practiced in adults, consciously creating an intense urge to initiate smoking in adolescents and young adults of 15 and 24 years as the most vulnerable groups [5-8]. Concurrent evidence indicates that the first instance of smoking most frequently occurs by the age of 18 [9,10], and with its reinforcing properties, continues up to young adulthood (18–24 years), which accounts for the highest cigarette intake rate compared to any other age groups [11,12].
Nonsmoking adolescents have also reported experiencing feelings of smoking, even with minimal or no history of tobacco exposure, thus leading to a degree of vulnerability to their well-being [13]. For example, DiFranza et al. [14] found that youths who have never used tobacco reported a feeling of being inclined to cigarettes, and many expressed a strong desire to start smoking. Initiating smoking at a very young age exposes individuals to negative physical health consequences, including heart disease [15], respiratory illnesses [16], cancer [17], and so on. Smoking initiation also places adolescents at risk of several adverse outcomes, such as looking older earlier than normal, being addicted to alcohol or other illicit drugs, and, more importantly, a deterioration in psychological well-being [18-20]. In past years, Bangladesh was one of the top ten countries in the world to use tobacco publicly in both smoking and smokeless forms due to the frequent availability of tobacco products, lack of proper tobacco control policies, and poor enforcement of the regulations [21,22]. Therefore, along with robust tobacco control strategies and effective prevention strategies targeting smokers, prevention of never-smoking youth from susceptibility to smoking and the initiation of smoking behaviors is important [23].
Smoking susceptibility is a cognitive predisposition characterized by a lack of commitment to disengage from cigarette use, ultimately reflecting an inclination to potentially engage in experimental smoking rather than conventional smoking behavior [24,25]. Several factors have been identified as potentially associated with such susceptibility in adolescents and young adults. Understanding and characterizing them is essential as prevention and intervention programs should be customized based on the specific risk and protective factors targeting such groups at different smoking stages, such as preparation, initiation, experimentation, regular use, and addiction [26,27]. Factors influencing adolescents and young adults toward smoking initiation have been explained from the interpersonal, societal, and socio-cultural contexts based on the Theory of Triadic Influence [28]. From the interpersonal viewpoint, those who express higher susceptibility to smoking have lower school attachment, sloppy self-efficacy, degraded self-esteem, and devalued self-image [13,29,30]. The social contextual factors associated with smoking susceptibility include addictive close company, smoking family members, and residing in a neighborhood having little or no control over second-hand smoke exposure [29,31,32]. Broader socio-cultural contextual factors are school-based smoking policies, the prevalence of tobacco retailers in proximity to school areas and attending places with significant tobacco use rates [32-35].
Considering these perspectives, the creation of awareness about the negative consequences of smoking in young adults to motivate them not to initiate it is crucial [35]. Concordantly, Hu et al. [36] identified the impact of anti-smoking media campaigns on cigarette usage and revealed that from 1989 to 1992, media campaigns led to a reduction in per capita cigarette consumption to 7.7 packs per 100 packs. Mattila et al. [37] documented that anti-smoking advertisements with emotional appeals generated higher awareness compared to rational and simply informative advertisements. Studies that established that anti-tobacco advertising with higher awareness and emotional contents exerted positive anti-smoking effects in adolescents and young children were consistent with this finding [38]. Thus, anti-tobacco awareness campaigns with emotional appeals arguably contribute to young adults’ decisions to remain smoke free [39]. Drawing from the existing literature [40,41], the tobacco industry’s promotional campaigns, inadequate anti-smoking education, lower exposure to anti-smoking media messages, lack of awareness, and poor implementation of tobacco control laws contribute significantly to creating strong desires toward smoking initiation in young adults. Despite this fact, well-validated tools for assessing adolescents and young adults’ smoking susceptibility and anti-smoking awareness have yet to be developed. Most studies measuring smoking susceptibility and anti-smoking awareness targeting such groups are primarily survey-based. From this viewpoint, to encourage the never-smoking youth not to initiate smoking by boosting the anti-smoking awareness program worldwide, based on a sample of adolescents and young adults in Bangladesh, the present study developed and validated the Smoking Susceptibility Measure (SSM) and the Anti-Smoking Awareness Scale (ASAS).
Accordingly, the study goals were: 1) to determine the content validity and internal consistency reliabilities of the SSM and ASAS; 2) to explore the factor structures of the intended measures using exploratory factor analyses (EFA), and confirm factor retention using parallel analyses; 3) to validate the factor structures through confirmatory factor analyses (CFA); 4) to investigate the convergent validity of the measures by determining the relationships of SSM and ASAS with school connectedness; 5) to evaluate whether SSM and ASAS can be used across age and sex; and 6) to demonstrate the contributions of anti-smoking awareness, school connectedness, and self-esteem regarding smoking susceptibility of adolescents and young adults.
METHODS
Participants
A total of 312 adolescents and young adults (13 and 23 years [mean=18.42±2.31]) recruited via convenience sampling from different schools, colleges, and universities of Rajshahi Division, Bangladesh, participated. In respect of demographic diversities among participants, convenient sampling technique was employed, reflecting an approximately balanced distribution for age (12–18 years adolescents/19–24 years young adults) and sex (males/females), but an imbalanced distribution for parental occupational status. Participants were recruited based on the criteria that they have never smoked, which was confirmed through their “no” responses when asked whether they had ever tried or experimented with cigarette smoking, even one or two puffs. For adolescents (12–18 years), students at high schools and colleges from 8th to 12th grade (n=179, 57.37%) and for young adults (19–24 years), undergraduate students from 1st to 3rd year (n=133, 42.63%) provided their responses with prior consent to participate. For adolescents aged 12–18, informed consent was also obtained from their parents and teachers (e.g., in cases where parents were absent). Regarding parental occupational status, 32.05% of fathers were businessmen (n=100), followed by non-government service holders (n=73, 23.40%), teachers (n=59, 18.91%), farmers (n=46, 14.74%), and government service holders (n=25, 8.01%). Mothers’ occupations were predominantly housewives (n=270, 86.54%), although a few were service holders (n=42, 13.46%). Among the participants, the percentage of males was 48.08% (n 1=150) whereas, the percentage of females (n 2=162) was 51.92%. The ethical approval of the research was taken from the Ethical Review Committee-Research and Publication (ERCRP), Department of Psychology, University of Rajshahi, Rajshahi-6205, Bangladesh [approval code: ERCRP-PSYRU-3(2)24; Date: 14.01.2024].
Procedure
Previous studies measuring the “smoking susceptibility” and “anti-smoking awareness” of adolescents and young adults rarely provide an integrated overview of each of the constructs through any valid instruments. To address this, we developed and validated two scales, namely, SSM and ASAS, following the guidelines on the construction of psychological instruments [42]. First, we defined the two constructs smoking susceptibility and anti-smoking awareness, and an initial item pool of 12 items for SSM and 10 for ASAS was then generated after reviewing the relevant literature [43-45]. Cognitive interviews with participants representing the target population were also conducted during the generation of items. While identifying the initial item pool of both SSM and ASAS, we considered several relevant aspects of smoking susceptibility and anti-smoking awareness based on the above-mentioned literature within a broad range of socio-cultural contexts.
The scales were developed under the following framework: 1) defining the purpose of the instrument; 2) generating the items; 3) reducing the items; 4) psychometric tests; and 5) deciding the final revision of the instrument. For SSM, we ensured that the various aspects influencing smoking susceptibility in adolescents and young adults have been covered. These aspects include participants’ level of interest in future smoking, exposure to smoking at home by parents, siblings, and senior family members, prior experience of taking one or two puffs, discussion among classmates on the attractive aspects of smoking, and so on. For ASAS, items were generated focusing on aspects such as banning smoking in public places, psycho-physical complications that smoking creates, exposure to anti-smoking media messages, prohibition of smoking at social gatherings, school education about the potential dangers of smoking, and so on.
Once the initial pool and formatting of items were accomplished for both SSM and ASAS, the items were reviewed, tested, and refined in multiple steps. In the first step, items were edited and revised based on cognitive interviews conducted with 1st to 3rd-year undergraduate bilingual students (n=10). Cognitive interviews were utilized to evaluate participants’ understanding and interpretation of items in both Bangla (native) and English versions. The participants agreed on the comprehensibility and interpretation of the items measuring SSM and ASAS. The linguistic equivalence of the Bangla and English versions of the intended measures was determined through the consensus of the interview participants [46]. In this way, both measures were finalized for the pilot study. Responses in both SSM and ASAS were recorded in a five-point Likert format where the scores ranged from 1 “very unlikely” to 4 “very likely” for each positively worded item. The scoring pattern was reversed for the negatively worded items. After cognitive interviewing, the items were reviewed by a panel of experts (2 Psychologists, 1 Clinical Psychologist, and 2 Associate Professors) with experience in the construction and adaptation of psychological instruments. They judged the representativeness, relevance, and clarity of the items on a 5-point response format ranging from 1 “least” to 5 “most.” Based on the expert recommendations and agreement scores, four and three items from SSM and ASAS were deleted, respectively, and the others were edited and/or revised. To verify the content expert validation, the judgment quantification by Content Validity Coefficient (CVCj) [47] was computed for each item (CVCi). For the total items of each scale (CVCt), CVC values ≥0.80 were used as the criterion for content validity [48]. Accordingly, eight and seven items for SSM and ASAS were retained for the final versions of the scales, respectively. The study was conducted following the ethical principles of research with human subjects [49], and ethical approval was obtained from the appropriate ethical review board.
Measures
SSM
The SSM consisted of eight items scored in a Likert-type response format ranging from very unlikely to very likely. Among the eight items, three were reverse coded (items 2, 7, and 8; 4=very unlikely, 1=very likely). For the remaining items (items 1, 3, 4, 5, and 6), the scoring pattern ranged from 1 to 4 (1=very unlikely, 2=rather unlikely, 3=somewhat likely, and 4=very likely). Higher scores indicated higher smoking susceptibility, whereas lower scores indicated lower susceptibility.
ASAS
The ASAS was comprised of seven items to measure the anti-smoking awareness of adolescents and young adults. Responses were recorded on a 4-point Likert scale ranging from very unlikely to very likely. Some items (items 3, 4, and 6) were reverse coded, with the scoring pattern for the reverse coded items ranging from 4 “very unlikely” to 1 “very likely.” For the remaining items (items 1, 2, 5, and 7), participants’ responses were scored from 1 “very unlikely” to 4 “very likely.” Higher scores meant higher awareness, whereas lower scores indicated lower awareness.
School Connectedness Measure
The Bangla translation of the 5-item School Connectedness Measure (SCM) adopted from Kaai et al. [27] was used to measure participants’ attachment to school. The responses were recorded on a 4-point Likert scale ranging from 4 “strongly agree” to 1 “strongly disagree.” Higher scores represented greater connectedness to school, whereas lower scores indicated lower connectedness. The internal consistency reliabilities of the SCM were within the acceptable ranges (ω=0.67, α=0.66).
Global Self-Esteem Measure
Consisting of 3-items, the Global Self-Esteem Measure (GSEM; based on the Self-Esteem Scale, Rosenberg [50]) was developed by Kaai et al. [27] The Bangla translation of this scale was used to measure participants’ global self-esteem. The responses ranged between true and false (true=4, mostly true=3, sometimes true/sometimes false=2, mostly false=1, and false=0). Higher scores indicated higher self-esteem, whereas lower scores indicated lower self-esteem. The internal consistency reliabilities of the translated GSEM fall within the acceptable ranges (ω=0.67, α=0.64).
Data analyses
Data were processed and analyzed using IBM SPSS (version 26, IBM Corp.), RStudio (version 2023.12.1.402, Posit), and Microsoft Excel 365. The psychometric properties of the SSM and ASAS were assessed using the Classical Test Theory (CTT) and advanced psychometric approaches.
CVC were determined for both SSM and ASAS. The normality of the data was checked through skewness and kurtosis. A skewness value of <3 and a kurtosis value of <10 were considered indicators of data normality for a large sample size (n>300 51 ). In CTT, mean inter-item correlations (ranged between 0.15 and 0.50) [52], corrected item-total correlations (accepted value≥0.30 [53]), internal consistency reliabilities (Cronbach’s alpha and McDonald’s omega, accepted reliability≥0.70) [54], EFA, parallel analyses, CFA, and multi-group CFA were utilized. Factor loadings explored in CFA were used to calculate the average variance extraction (AVE) (accepted value≥0.50 [55]) and the composite reliability (accepted coefficient≥0.70 [55]). Before running EFA, determinant value (>0.0001 [56]), Kaiser–Meyer–Olkin (KMO) (>0.60 [57]), and Bartlett’s test of sphericity (p<0.001 [57]) values were checked to support the adequacy of the data for EFA [53]. In EFA, for the retention of factors, the recommended eigenvalue was considered ≥1 (the Kaiser–Guttman criterion [58]). A parallel analysis was conducted to confirm the factor retention in EFA. In CFA, model fits were assessed through the indices of χ2/df (<5 [59]), comparative fit index (CFI), Tucker–Lewis index (TLI) (≥0.90 [60]), and root mean square error of approximation (RMSEA) (≤0.08 [61]).
Multi-group CFA was performed to assess the invariances of the SSM and ASAS across sex and age. The values of ΔCFI ≤0.010 and ΔRMSEA≤0.015 were considered as the indicators of measurement invariance [62]. Although a non-significant value of Δχ2 indicates invariances along with ΔCFI and ΔRMSEA, we did not consider Δχ2 as an indicator as it is a rigorous and sample-sensitive test of invariance [63].
Multiple hierarchical regression was conducted to investigate the contributions of adolescents’ and young adults’ anti-smoking awareness, school connectedness, and self-esteem on smoking susceptibility.
RESULTS
Baseline demographic characteristics were presented in Table 1. For descriptive statistics (Table 2), the means and standard deviations (SDs) of the individual items ranged between 1.26 (SD=0.61) and 1.94 (SD=0.69) for SSM and 2.37 (SD=0.73) and 3.51 (SD=0.77) for ASAS. The item-level information indicated that the skewness (0.44 to 2.64 for SSM, -1.52 to 0.25 for ASAS) and kurtosis (0.35 to 6.77 for SSM, -0.58 to 1.21 for ASAS) values fall within the suggested limits (<3 for skewness and <10 for kurtosis [49]). Furthermore, all items had good corrected-item-total correlations, which ranged between 0.49 (item 8) and 0.80 (item 7) for SSM and 0.48 (item 7) and 0.71 (item 2) for ASAS (Table 2). The mean item-total correlations were 0.52 for SSM and 0.43 for ASAS.
Baseline demographic characteristics of the participants (N=312)
Item-level psychometric properties of the SSM and ASAS
The internal consistency reliabilities of both scales reached an excellent level, as demonstrated through Cronbach’s alphas (0.90 for SSM, 0.84 for ASAS) and McDonald’s omegas (0.90 for SSM, 0.85 for ASAS) (Table 3). The composite reliabilities (0.89 for SSM, 0.84 for ASAS) of the newly developed measures were in line with the suggested limits.
Scale-level psychometric properties of the SSM and ASAS
EFA was conducted to determine the factor structure of SSM and ASAS. Prior to this, KMO values (0.85 for SSM, 0.86 for ASAS) and Bartlett’s tests of sphericity values (2,725.02, p<0.001 for SSM; 1,150.70, p<0.001 for ASAS) suggested the suitability of factorizing the sample through EFA. The results of EFA revealed a single-factor structure for both measures, cumulatively explaining 67.90% and 52.20% of the variances in SSM and ASAS, respectively. The EFA was conducted using the minimal residual approach with oblique rotation (geominQ). To confirm the factor retention, a parallel analysis was run, which established greater real data component mean eigenvalues (5.72 for SSM and 4.09 for ASAS) than those of the simulated data (1.26 for SSM and 1.23 for ASAS) for the single factor structure of both measures.
To validate the factor structure of SSM and ASAS, we conducted CFA (Table 4) using the diagonally weighted least square method. The following indices were obtained from the CFA using the cut-offs for the fit criteria; the single-factor structure of the SSM and ASAS fitted the data well (χ2/df=5.24, CFI=0.99, TLI=0.99, RMSEA=0.11 for SSM; χ2/df=3.17, CFI=0.99, TLI=0.99, RMSEA=0.08 for ASAS) (Table 4). For SSM, EFA factor loadings of items ranged between 0.60 (item 8: second-hand smoke exposure in school areas or public places does not create an urge to smoke) and 0.92 (item 1: I might try smoking cigarettes in the future; item 2: regular smoking of siblings or same age group family members does not trigger me to smoke), and those for CFA ranged between 0.59 (item 8) and 0.94 (item 1). For ASAS, EFA factor loadings ranged from 0.56 (item 7: in co- and extra-curricular activities, we are taught to be aware of the negative consequences of smoking) to 0.87 (item 1: I am in favor of banning smoking in public places), and CFA factor loadings ranged between 0.56 (item 7) and 0.91 (item 2: I firmly believe that cigarette smoking is harmful to health) (Table 5).
Model fit indices of SSM and ASAS using confirmatory factor analysis
Factor loadings of the items of SSM and ASAS
A multi-group CFA was performed to test the invariability of the measures across age and sex. The values of ΔCFI and ΔRMSEA presented in Table 5 fall within the recommended ranges, which indicate that the SSM and ASAS can invariably be applied to males and females based on sex as well as on adolescents and young adults based on age (Table 6).
Measurement invariances of the SSM and ASAS across age and sex
While considering the convergent validity of SSM and ASAS, school connectedness was used as the criterion variable to determine the correlation coefficients with newly developed scales. A significant negative correlation between the total score of SSM and that of SCM revealed that the higher adolescents’ and young adults’ school attachment, the lower their smoking susceptibility and vice versa (r=-0.23; p<0.001; 95% confidence interval [CI], [-0.12, -0.33]). By contrast, a significantly positive association of school connectedness with the ASAS suggested that the higher the school attachment, the higher the awareness regarding the negative aspects of smoking and vice versa (r=0.26; p<0.001; 95% CI, [0.15, 0.36]). The coefficients indicated the convergent validity of the newly developed measures. Moreover, the AVE values (0.72 for SSM, 0.54 for ASAS) extracted from the CFA factor loadings were in line with the suggested ranges, also reflecting the convergent validity of the SSM and ASAS.
Hierarchical regression analysis (Table 7) indicated that at stage one, anti-smoking awareness significantly contributed to the model (F [1, 310]=21.41, p<0.001) and accounted for 6.5% of the variations in smoking susceptibility. The inclusion of the school connectedness variable at stage two explained a further 2.9% of the variation in smoking susceptibility, and this change in R2 was significant (F [1, 309]=9.76, p<0.01). When all three predictors (i.e., anti-smoking awareness, school connectedness, and self-esteem) were entered in stage three of the regression model, self-esteem uniquely accounted for a 1.9% change in smoking susceptibility, and this change appeared significant (F [1, 308]=6.46, p<0.05). More specifically, for all three predictors in stage three, smoking susceptibility was negatively predicted by anti-smoking awareness (β=-0.19, t=-3.40, p<0.01) and school connectedness (β=-0.14, t=-2.53, p<0.05) but positively predicted by self-esteem (β=0.14, t=2.54, p<0.05), thereby indicating higher anti-smoking awareness and greater school connectedness lowered the risks of smoking susceptibility. Interestingly, adolescents and young adults who scored higher in self-esteem exhibited higher smoking susceptibility.
Hierarchical regression of smoking susceptibility based on anti-smoking awareness, school connectedness, and self-esteem of adolescents and young adults
DISCUSSION
Although smoking susceptibility in never-smoking youths warrants global consideration, the majority of research has been conducted in higher-income countries compared to lower-middle-income countries [23,32]. Concordantly, some frequently used smoking susceptibility measures [31] followed the strictly defined criterion by ignoring potential contributing factors, thus restricting their primary outcomes to differentiate susceptible groups from non-susceptible ones. By contrast, those that focused on contributing factors associated with the transition from smoking experimentation to initiation in adolescents and youths were primarily survey-based [64]. Different socio-cultural factors including peer influence, curiosity, smoking family members, and social media exposure of tobacco promoting contents are prevailing in communities of developing countries including Bangladesh for smoking susceptibility and initiation in adolescent and young adult students [21,65-68]. Particularly, social media and peer group influences for smoking susceptibility and initiation among students are most prevalent in urban areas where social and family customs create perception among young adults to consider smoking as a normal behavior despite increased awareness about the harmful effects of smoking [69-78]. In this respect, the present study sample consisted of adolescent and young adult students recruited from the institutions of urban areas in Bangladesh. The socio-cultural norms about smoking in Bangladesh promote the tobacco industry to market their products intensively and therefore, young adults get cigarettes from the sellers without any age restrictions [79]. Taken together, the relationship between susceptibility and smoking initiation evidently depends on factors including family and peer smoking status, school culture and attachment, second-hand smoke exposure, availability and uses of other substances, and the mental states of never-smoking youths with variations in sex, school grade, culture, and ethnicity. In past years, smoking was more prevalent in males due to social, cultural, and religious restrictions of smoking in females. This trend has recently changed, exhibiting a strong attraction to initiate smoking or a rising trend of smoking in females [21,22,79,80]. However, the smoking prevalence rate in recent years has been reduced significantly to 18% in Bangladesh due to the strong initiatives of the government like policy implementation, increased taxation, and community-based interventions [81-85]. Bangladesh is trying to strictly implement the national tobacco control laws and regulations and, imposing duty and VAT on the unmanufactured tobacco products, incorporating provisions of the WHO Framework Convention [79]. Moreover, to decrease the youth susceptibility to smoking and prevent transition to future regular smokers, anti-smoking awareness programs highlighting the harmful effects of smoking are providing protection against smoking initiation and increasing the propensity to quit smoking [32].
Altogether, national tobacco control plans, preventing second-hand smoke exposure in the community [64], empowering youths to resist peer and sibling pressures [86], school-based anti-smoking education [64], school-level tobacco control strategies [87], and restricting media campaigns on tobacco promotions [32] have the capacity to reduce never-smoking youths’ susceptibility to initiate smoking. Considering this aspect, the present study addressed the interpersonal and social context influences of smoking susceptibility and anti-smoking awareness in adolescents and youths and developed two new scales, SSM and ASAS, covering the potential contributing factors discussed herein.
Relevant literature reviews, expert recommendations, and agreement scores of the pilot testing resulted in the retention of eight and seven items for SSM and ASAS, respectively. The newly developed SSM and ASAS were administered to a sample of 312 adolescents and youths to determine further psychometric properties. Prior to this, the normality of the data was checked and confirmed through skewness and kurtosis coefficients of less than three and less than ten, respectively [51]. Both rating scales had good internal consistency reliabilities [54], demonstrated through Cronbach’s alphas and McDonald’s omegas. Furthermore, item-level psychometric properties such as mean inter-item correlations, corrected item-total correlations, composite reliabilities, and AVE values reached the suggested limits. To determine the factor structure of the SSM and ASAS, we conducted EFA and parallel analyses to confirm the factor retention. Prior to this, KMO and Bartlett’s test of sphericity values proved the adequacy of the data to factorize the sample. EFA results produced a single-factor structure for both measures, and the same single-factor structure was retained in parallel analyses. To assess the structural validity of both rating scales, CFA was performed and had an adequate model fit to the data for the unifactorial structure of SSM and ASAS.
We conducted multi-group CFA on the present data to check whether both scales can assess smoking susceptibility and anti-smoking awareness in the same way across sex and age. From the analyses, both scales can evidently be applied across adolescents and young adults in respect of age and to males and females regarding sex. Previous findings revealed that males consistently outnumbered females in the high smoking susceptibility group, and they were more likely to be regular smokers as opposed to experimental smokers over time. Males were found to be protected from such tendencies with higher resilience scores, whereas females have a positive peer group. Moreover, previous studies also revealed that with each passing year, students had a higher susceptibility to smoking initiation, which suggests that smoking prevention strategies should preferably be implemented early when students are in lower grades [27]. These findings clearly indicated the importance of sex and age to check whether the SSM and ASAS can be applied across these variables for measuring participants’ smoking susceptibility and anti-smoking awareness, focusing special attention on sex and age in developing interventions [88].
However, while determining the convergent validity of the SSM and ASAS, participants’ school connectedness was considered as the criterion variable assessed through the SCM [27]. Higher school connectedness was associated with lower smoking susceptibility and higher anti-smoking awareness. The findings that students’ lower smoking susceptibility and reduced smoking initiation were evident when they feel more connected to school, thereby creating a school culture that gives students a sense of belonging, makes them skilled to resist tobacco use and connects them with community resources to intervene their smoking family members and peers [29,87]. To this end, community-based programs led by community leaders such as religious persons, teachers, youth, and civil society can strengthen school-based tobacco control programs and policies [79]. Moreover, media advocacy intended to reduce tobacco advertisements and promotions in social media and in-person peer gatherings should be programmed to discourage never-smoking youths’ smoking initiation by warning them about the health hazards of smoking, ultimately promoting a smoke-free school culture [32,85,89,90].
Finally, while exploring the contributions of anti-smoking awareness, school connectedness, and self-esteem in predicting smoking susceptibility of adolescents and young adults, smoking susceptibility was predicted by low levels of anti-smoking awareness, low levels of school connectedness, and high levels of self-esteem. Consistently, never-smoking youth who were exposed to anti-smoking awareness programs and school-based anti-smoking education demonstrated decreased susceptibility to smoking [4]. While considering school connectedness as a predictor of smoking susceptibility in never-smoking youth, the strength of school-based tobacco control programs and policies was found to create a sense of belongingness, ultimately working as a protective factor for smoking susceptibility [89]. In terms of self-esteem, contrary to the previous studies in which low self-esteem resulted in self-dissatisfaction, self-rejection, and predisposed the never-smoking adolescents and youths toward high smoking susceptibility [91], we observed a positive association between self-esteem and smoking susceptibility. However, the underlying reasons behind our findings may be that adolescent smokers develop an identification as a smoker that creates a perception of normality and social acceptance, thus reflecting maturity, independence, and enhanced self-image [91]. In school contexts, adolescents and youths feel a desire to be socially/culturally valued by their peers and adopt smoking as a maladaptive means of coping, ultimately attempting to bolster self-esteem through an alternate source of identity [92]. In Bangladesh, the smokers perceive smoking as a trend for individuals in higher social status, thinking that smoking can make them look smarter and more modern compared to non-smokers [21]. While mixing with the smoking peers, this misperception is transferred to the never-smoking youths and creates an urge to initiate smoking for a smarter social appearance and acceptance [66-68,93-96], leading to enhanced self-esteem.
Limitations and future suggestions
The study has the following limitations. First, the data were cross-sectional and represented adolescents and young adults from a divisional town in Bangladesh. The geographic or cultural homogeneity in selecting participants of the present study limits the generalizability of the findings across cultures. Next, the participant’s responses in SSM and ASAS were subjected to self-report bias whereby participants may overestimate the display of maturity through smoking initiation or underestimate by demonstrating lower smoking susceptibility to please the experimenter [97]. However, confidentiality of responses was ensured through the informed consent of the participants and/or their caregivers, and therefore, self-report bias did not compromise the integrity of the study. To the best of our knowledge, the present study is the first approach to develop and validate two rating scales in a sample of Bangladeshi adolescents and young adults. Future research should validate these measures in different cultures and across varied populations covering students of primary, secondary, and higher secondary levels of education. Taken together, the newly developed SSM and ASAS can help psychologists, educators, school staff, teachers, and significant others to identify adolescents and youths at risk for smoking before initiation, thus leading to appropriate intervention strategies in reducing the number who become regular smokers.
Notes
Availability of Data and Material
The datasets generated or analyzed during the study are available from the corresponding author on reasonable request.
Conflicts of Interest
Seockhoon Chung, a contributing editor of the Psychiatry Investigation, was not involved in the editorial evaluation or decision to publish this article. Mohd. Ashik Shahrier has declared no conflicts of interest.
Author Contributions
Conceptualization: Mohd. Ashik Shahrier, Seockhoon Chung. Data curation: Mohd. Ashik Shahrier. Formal analysis: Mohd. Ashik Shahrier. Methodology: Mohd. Ashik Shahrier, Seockhoon Chung. Writing—original draft: Mohd. Ashik Shahrier. Writing—review & editing: Mohd. Ashik Shahrier, Seockhoon Chung.
Funding Statement
None
Acknowledgements
None