Association for Public Policy Analysis & Management

Introduction: The Supplemental Nutrition Assistance Program (SNAP, formerly known as the Food Stamp Program) is the federal government’s primary form of food assistance to lower-income Americans, and is one of the largest social safety net programs, costing over $70 billion in 2015. The modern program, which serves over 46 million individuals, includes means-testing and work requirements, which have been the topic of Congressional debate and reform efforts since its formalization under the federal Food Stamp Act of 1964. Since a large program expansion under the American Recovery and Reinvestment Act, the size, scope, and nature of SNAP has been increasingly scrutinized by elected officials. Concerns over program fraud, unhealthy eating, and luxury purchasing have also punctuated recent discourses.

Previous studies have demonstrated how through the selection and framing of topics, the news media acts as an agenda-setter - impacting the salience of public issues and influencing public opinion and policy-makers alike. These sorts of analyses have often included a sample of articles published in major newspapers; however, with the advent of online journalism and social media, an increasing number of individuals access news via digital platforms. Thus, a much broader analysis - requiring new analytical tools - is needed to understand the full scope of coverage regarding specific programs. This study aims to quantitatively and qualitatively assess when, where, and how SNAP is covered by newsmedia in the US.

Two primary research questions drive this study: 1) How have major newsmedia outlets (e.g., New York Times, Wall Street Journal, etc.) characterized SNAP over time, and 2) How has online media influenced these dynamics?

Methods: We used a Google News web-scraping algorithm to generate a database of titles, dates, sources, and complete texts of all items published since 1990 that were returned by searching “Supplemental Nutritional Assistance Program,” or “Food Stamp;” to help reduce the number of false-positives, a text classifier (logistic regression implemented in Scikit learn) was used. We performed preliminary topic modelling with an open-source machine learning program, MALLET (Machine Learning for Language Toolkit), to identify themes of reporting about SNAP. Additional data-mining procedures will be introduced to ensure that older documents are appropriately represented in the final database. Descriptive statistics will be performed to assess changes over time and within/between outlet types. With these themes, we will also select a random sample of ten articles from the top five themes for qualitative coding, both as a check for the validity of the topic modelling, and to further analyze the nature of the discourse around SNAP.

Preliminary Findings: The database assembled via web-scraping algorithm included over 8,200 articles identified as being about SNAP. The top five thematic areas within this preliminary database include: legal and administrative issues, health and nutrition, poverty, children and families, and fraud. We are generating a final database and investigating the distribution of thematic areas by traditional media sources vs. online sources and how this changes over time, and will begin the qualitative coding portion of the study during late Spring 2017.