Toward equitable science communication in an algorithmically infused society: Understanding media and message antecedents to knowledge of wicked science issues
In an era of increasingly fast-moving and disruptive scientific and technological advances, communicating science to broad segments of the population becomes more important than ever. The persistent challenge to reach diverse population segments, especially those that are traditionally underserved by science outreach efforts, is complicated by the fast-evolving information environment we live in. Many of today’s media platforms are driven by artificial intelligence (AI) algorithms, which tailor information to our personal preferences, biases, and contexts. Despite the deepening integration of these algorithmically driven media tools—such as social media platforms and online news aggregators—into our everyday life, it remains unclear how these media tools distribute scientific information in society across diverse population segments. It is also unclear how such media infrastructures might shape important individual and collective outcomes, such as knowledge of emerging science and technologies. Moreover, research and practice should continue to develop communication strategies for addressing information inequity in contemporary media environments. This dissertation investigates how the media and their message-level factors might affect diverse social segments’ knowledge of three wicked science issues, namely human gene editing, artificial intelligence, and COVID-19 and its related vaccines, in today’s algorithmically infused information environment. It begins by reviewing the current algorithmically infused information environment, explain disparities in science knowledge and how the new information environment shapes them, and comparing the three wicked science issues that make up the contexts of inquiry. This dissertation then uses a sock puppet algorithm audit design to explore the extent to which algorithmically driven social media platforms such as YouTube recommend science content based on users’ racial and socioeconomic status (SES) profiles (Study A). From there, this dissertation uses three public opinion survey datasets to examine whether use of different social media platforms affects the gaps in factual and perceived knowledge of wicked science issues among Americans with different racial and SES makeup (Study B). Finally, using an experiment, this dissertation examines how message characteristics such as information modality (visual versus text-based) and rhetorical mode (narrative versus logical-scientific) could be leveraged against inequalities in understanding of and engagement with health science information among individuals with varying levels of science literacy (Study C). Results show that social media algorithms, such as the YouTube algorithm, can indeed expose sociodemographically diverse audiences to different subsets of information even when people are actively searching for the same science issues, although the degree of information tailoring may depend on the specific search topic (e.g., how heavily the issue is discussed by different sources on YouTube). Higher-SES audiences, especially higher-SES White American audiences, are likely to receive a wider range of video and channel recommendations when searching for science issues than lower-SES audiences. In addition, use of different social media platforms is overall associated with wider gaps in factual knowledge of the three science issues and narrower gaps in perceived knowledge of the three wicked science issues among racial minorities than among Whites. Finally, exposure to visual or narrative messages about COVID-19 vaccine safety significantly reduces the gap in factual knowledge of COVID-19 vaccine safety between high and low science literacy groups, while exposure to narrative (but not visual) messages also reduces the gap in message elaboration between the groups. However, combining visuals and narratives does not further enhance message effectiveness. The theoretical, methodological, and practical implications of this dissertation are discussed.