Smoking and self-perception in secondary school students

Abstract

Correspondence to : Dr Anne Charlton, Director, Cancer Research Campaign Education and Child Studies Research Group, Department of Public Health and Epidemiology, Stopford Building, University of Manchester, Oxford Road, Manchester, M13 9PT, UK Abstract Objective To carry out a cross-sectional pilot baseline survey to investigate poss ible links between teenage smoking and self-perception in six domains by means of a validated psychological measure. Design Anonymous self-administered questionnaires including the Harter Self Perception Profile for Children (1985). Subjects Eight hundred and thirty students (407 boys and 423 girls) aged 11 to 15 years in two state secondary schools in Northern England. Main outcome measures Chi-squared tests, odds ratios, and one-way analysis of variance for age trends with regard to self-perception and smoking. Correlation between domains, and regression analy sis with domains and with smoking as the dependent variable. Results -The girls' self-perception was significantly lower than that of the boys for scholastic, athletic, physical appear ance, and global self-worth domains. Girls' self-perception scores fell signi ficantly with increasing age as follows: scholastic (p = 0.0300), physical appear ance (p < 0.0001), behavioural (p < 0.0001) and global self-worth (p < 0.0001) domains, whilst boys showed no signi ficant changes. Smoking was significantly linked to low self-perception scores in girls in scholastic competence (p < 0.0001), physical appearance (p = 0.0011), behavioural conduct (p < 0.0001), and global self-worth domains (p < 0.0001). For boys, only in scholastic (p = 0.0010) and behavioural (p < 0.0001) self-percep tion was smoking linked with low scores. Higher self-perception scores in the social competence domain were associated with smoking in boys and girls, but not signi ficantly. Associations were found between most of the six domains, the most signi ficant being between physical appearance and global self-worth for girls, and the only pair that showed no association for either boys or girls was behavioural conduct and social acceptance. Logistic regression analysis showed a combination of friends' smoking, school attended, and self-perception with regard to behaviour and social acceptance to be the most significant factors associated with smok ing in girls, with male family members' smoking being significant for boys rather than school or social acceptance. Outcome-A longitudinal study is to be conducted to investigate possible caus ality. The meaning of the findings for the health-promoting school are discussed.