This study is an analysis of adolescent risk behaviors contributing to an increased rate of suicidal ideation for 12 to 18 year olds. The Youth Risk Behavior Surveillance System Survey (YRBSS) is an epidemiologic survey designed to monitor the prevalence of risky behaviors of adolescents in middle and high school1. The YRBSS is a complex sample survey with a three-stage cluster design. Multiple logistic regression is used to analyze the data, including methods of analysis to address issues in complex survey design. Results of this study indicate several different risk factors that influence the rate of suicidal ideation among adolescents, including alcohol and drug use, sexual risky behaviors, unhealthy weight loss methods, depressed mood, sex and race/ethnicity. The conclusions of this study indicate that many risk factors associated with suicidal ideation are behaviors that could be addressed with early intervention strategies to reduce the risk of suicidal ideation.

[pt] Esta dissertação avalia o impacto do nível educacional dos brasileiros, na sua remuneração, levando-se em conta o grupo da posição na ocupação por ele exercida. A análise é focalizada no universo de indivíduos ocupados com mais de 10 anos de idade, moradores na área urbana do Brasil. Isto é feito considerando como base de dados a Pesquisa Nacional por Amostra de Domicílios - PNAD para os anos de 1992 e 1999. Por se tratar de uma pesquisa feita por amostragem estratificada de conglomerados com múltiplos estágios de seleção, não se pode assumir as hipóteses usuais de modelagem das observações, tais como o comportamento IID (independente e identicamente distribuídos) dos dados. Por este motivo a modelagem é feita utilizando o pacote estatístico SUDAAN que é apropriado para o tratamento de dados obtidos por amostras complexas. / [en] This dissertation evaluates the impact of the educational level of the brazilian works on their income, accounting for their position in the occupation. The analysis will be focused in the universe of employed people over ten years of age that leave in the urban area of Brazil. This will be done for the data set of the Brazilian Household Sample Survey (Pesquisa Nacional por Amostra de Domicílios - PNAD) for the years of 1992 and 1999. As the PNAD is a stratified multistage cluster sample survey, the usual hypothesis for modeling observations such as the IID data behavior, cannot be assumed. For this reason the modeling process is performed making use of the statistical software SUDAAN, which is adequate to treat data, obtained from sample surveys.