Analyzing Associations of Lifestyle Behaviors and Health-Related Variables Using Quantile Regression


This case study shares an experience using a specific analytical method to examine data from the 2017 U.S. Youth Risk Behavior Survey (YRBS). The study was a secondary data analysis of lifestyle behaviors and body mass index (BMI), which included a relatively underused quantile regression approach. After reading recent literature in the area of physical activity and health and how researchers are increasingly applying this analytical method to health outcome data, it was decided that the application of quantile regression to the nationally representative 2017 U.S. YRBS data will yield novel information of the relationship between specific self-reported lifestyle behaviors and BMI in adolescents. Use of quantile regression was appropriate because the distribution of BMI on the 2017 U.S. YRBS was negatively skewed and scores within the higher quantiles of BMI are clinically meaningful. Using the publicly available 2017 U.S. YRBS dataset and the accompanying Data User’s Guide, quantile regression was employed within a cross-sectional research design. The quantile regression approach was well received by the manuscript reviewers; however, it was also recommended that the mean-based linear regression coefficients also be computed and interpreted. Formatting of tables and figures to communicate the results of the quantile regression proved to be more difficult than reporting the results of mean-based regression only. Use of quantile regression is certainly appropriate when specific quantiles of a variable’s distribution are of clinical interest; however, knowledge of the meaning of scores within a variable’s distribution is needed when deciding whether or not to use this method.

locked icon

Sign in to access this content

Get a 30 day FREE TRIAL

  • Watch videos from a variety of sources bringing classroom topics to life
  • Read modern, diverse business cases
  • Explore hundreds of books and reference titles