Fatores associados a realização de exercício físico por mestrandos utilizando modelo de regressão de Poisson múltiplo e análise de correspondência múltipla

Abstract

Introduction: Physical exercise has been consolidating itself as a fundamentally important tool for improving health and quality of life. Objective: To compare the performance of statistical techniques, multiple Poisson regression model and multiple correspondence analysis, in the analysis of factors associated with physical exercise by master's students. Methodology: This is an observational, cross-sectional study, of the self-completed survey type, with data collection being carried out online, in a sample of 375 master's students who answered the Epworth Sleepiness Scale, SRQ-20 instruments and WHOQOL-bref, in addition to questions related to sociodemographic characteristics and lifestyle. Multiple Poisson regression model with robust variance was used to verify associations as well as multiple correspondence analysis. Results: Of the 375 participants, 57.34% claim to perform physical exercise, in relation to the characteristics and lifestyle of master's students, it was found that the variables alcohol consumption, satisfaction with health status and possibility of common mental disorder showed a significant association the performance of physical exercise, both through regression analysis and through multiple correspondence analysis. As for the facets of the quality of life measurement instrument, significant associations were obtained between performing physical exercise with energy and fatigue, body image and appearance, and opportunities to acquire new information and skills in the Poisson regression model. In the multiple correspondence analysis, only the last two remained in the final model. Conclusion: Both techniques proved to be effective in determining the associated factors, however, when there are many variables, the multiple Poisson regression model is simpler to perform.