Date of Award

6-6-2021

Document Type

Thesis

Publisher

Santa Clara : Santa Clara University, 2021.

Department

Computer Science and Engineering

First Advisor

Yuhong Liu

Abstract

With the outbreak of the COVID-19 pandemic, an overabundance of information related to the virus was released through social media. While more information can ultimately help people learn about the virus and how to protect themselves from it, there has been a lack of uniformity amongst individual states and the federal government in regards to policies and health guidelines (69), which may lead to the public being confused on the best guidelines to follow. There has also been an ideological divide driven by political polarization in the US on how to respond to this pandemic, causing greater contention (17; 86; 37; 46). Analyzing the public’s response to COVID-19 state-level policies on social media can help us to better understand the progression of the pandemic in the US.

We analyze public sentiment from Twitter, Facebook, and Reddit data related to policies such as shelter-in-place orders, fall school reopening guidelines, and face mask guidelines using machine learning sentiment analysis methods and compare the results by performing significance testing. We use the COVID-19 cases and deaths and state demographics to analyze and contextualize the results during the timelines that these social media posts were published. We found that users had more positive sentiments in response to mask policies than to shelter-in-place and fall school reopening policies and that there are statistically significant differences in public sentiment when comparing the same policies between several states. Finally, we found that when testing the significance of sentiment differences for different policies within the same state, the majority of the statistically significant differences were found within states that were considered to be swing states in 2020. We suggest that further research be done to analyze why these differences exist and what factors may impact public sentiment regarding COVID-19 public health policies.

Download

Share

COinS