Exploring the Predictors of Marijuana Use Among Adolescents with Asthma
Class
Article
Graduation Year
2018
College
Emma Eccles Jones College of Education and Human Services
Department
Psychology Department
Faculty Mentor
Ginger Lockhart
Presentation Type
Oral Presentation
Abstract
Regularized regression techniques are useful to analyze high-dimensional data common in psychology, prevention, and social sciences. Although currently unrecognized in prevention work, regularized regressions are extensively used in biology and genetics. Among these regression methods, the elastic net technique is especially flexible given it can handle large number of potentially correlated variables and can select the appropriate model to a high degree. These attributes make it particularly helpful in exploratory work.
We, therefore, present the advantages and disadvantages of elastic net and demonstrate its use on high-dimensional health-risk behavior data. The data are from the federally-funded Youth Risk-Behavior Surveillance System (YRBSS) from 2015. Due to the higher levels of acceptance of marijuana use among adolescents, we focus on assessing the predictors of marijuana use from a large list of health, demographic, and behavior indicators. Since marijuana use is especially dangerous among youth with asthma (as it is often smoked thereby increasing the risk of an asthma attack) we focus on answering the question, “In what ways may adolescents with asthma be at risk of using marijuana?” Results indicate that marijuana usage among adolescents with asthma is related to other problem behavior such as harder drug usage, violence, and other risky behaviors. The relationships found herein need to be assessed further in confirmatory analyses.
Location
Room 101
Start Date
4-13-2017 1:30 PM
End Date
4-13-2017 2:45 PM
Exploring the Predictors of Marijuana Use Among Adolescents with Asthma
Room 101
Regularized regression techniques are useful to analyze high-dimensional data common in psychology, prevention, and social sciences. Although currently unrecognized in prevention work, regularized regressions are extensively used in biology and genetics. Among these regression methods, the elastic net technique is especially flexible given it can handle large number of potentially correlated variables and can select the appropriate model to a high degree. These attributes make it particularly helpful in exploratory work.
We, therefore, present the advantages and disadvantages of elastic net and demonstrate its use on high-dimensional health-risk behavior data. The data are from the federally-funded Youth Risk-Behavior Surveillance System (YRBSS) from 2015. Due to the higher levels of acceptance of marijuana use among adolescents, we focus on assessing the predictors of marijuana use from a large list of health, demographic, and behavior indicators. Since marijuana use is especially dangerous among youth with asthma (as it is often smoked thereby increasing the risk of an asthma attack) we focus on answering the question, “In what ways may adolescents with asthma be at risk of using marijuana?” Results indicate that marijuana usage among adolescents with asthma is related to other problem behavior such as harder drug usage, violence, and other risky behaviors. The relationships found herein need to be assessed further in confirmatory analyses.