Latent Class Analysis for Classification of Latent Policy Environments: a Case Study

• Activity Number: 167 - SPEED: Missing Data and Causal Inference Methods, Part 1
• Type: Contributed
• Date/Time: Monday, July 29, 2019 : 10:30 AM to 12:20 PM
• Sponsor: Health Policy Statistics Section
• Author(s): Bryan Blette* and Leah Frerichs and Annie Green Howard
• Companies: University of North Carolina at Chapel Hill and University of North Carolina at Chapel Hill and The University of North Carolina at Chapel Hill
• Keywords: Latent Class Analysis; Policy Statistics; School Health; Obesity; Diet; Physical Activity

Abstract

Policy researchers are often interested in questions concerning an overall policy domain that encompasses many heterogeneous and unstandardized laws and policies. A question such as “How have climate change policies changed in industrialized nations over the last decade?” requires consideration of policies on energy efficiency, use of fossil fuels, tax credits, renewable energy, and others, inducing challenges in definition and measurement of the outcome of interest. Consequently, researchers frequently collect data on many outcomes and analyze each individually or combine variables into a composite outcome. Latent class analysis is a powerful tool for identifying unobservable subgroups in a population while performing a dimension reduction of variables of interest, but it has had limited use in policy research. We perform a case study using latent class analysis to derive one variable classifying the US high school food policy environment and one classifying the physical activity policy environment using CDC data. We investigate how the resulting latent class distributions changed from 2000-2016 and consider the advantages and limitations of using this method in policy research.