How the Media Shapes Public Perception: The Power of Agenda-Setting

• News coverage can both increase and decrease related discussions on social media, highlighting its powerful role in setting public agendas.
• Liberal and conservative media significantly influence social media discussions on immigration.
• Different social media platforms respond uniquely to news coverage and policy announcements.

Introduction

Ever wondered why certain issues dominate your social media feed? The answer lies in agenda-setting theory.

This is the main finding of a new study by Jisoo Kim of the University of Wisconsin, Yini Zhang of the University of Buffalo, and Porismita Borah of Washington State University. Published in Information, Communication & Society, a top communication journal, their study examines topics of discussion on social media and the influence of the media and policy in determining which topics proliferate and which do not.

Read on to learn more about their research into how the media sets the agenda for discussions on social media, and what you can do to navigate and critically engage with these discussions.

Background

Agenda-setting theory explores how the media influences public perception by highlighting certain issues, thereby making them more prominent in public discourse. This theory has been extensively tested and validated, with significant research focusing on how media coverage affects the public’s attention to specific issues.

A more advanced aspect of this theory, known as second-level agenda-setting, delves deeper into how the attributes of these issues are portrayed:

• Attributes are specific aspects or characteristics of a given issue. For example, when discussing political candidates, the media might highlight attributes such as morality or educational background.
• This portrayal can significantly shape public perception, making certain characteristics more noticeable and influential in how people view the issue. For instance, news coverage on immigration can emphasize different attributes like crime, economy, or terrorism, each shaping public discourse differently.

This study examines how social media platforms reflect public attention and opinion on issue attributes.

• Social media serves as an effective proxy for understanding public discourse, capturing real-time, naturally occurring opinions.
• For example, in debates about immigration, social media discussions can focus on either the border wall or family separation, influenced by ideological leanings and external events.
• This approach moves away from traditional reliance on surveys and seeks to understand public opinion as it organically manifests on platforms like Twitter, Facebook, and YouTube.

At the first level of agenda-setting, traditional media often drives the agenda, influencing digital platforms’ focus. However, political leanings within the media ecosystem play a crucial role. Liberal and conservative media channels cater to different audiences, reinforcing specific frames and issues within their partisan ecosystems. This selective consumption by partisans further shapes public discourse.

Public policy decisions also play a critical role in shaping public opinion and expression. Policymakers can influence media agendas, and in turn, public agendas. This influence is particularly strong on issues that undergo intense policy debates, such as immigration. During the Trump administration, policies like the “Muslim ban” and executive orders on immigration significantly influenced public discourse, especially during the COVID-19 pandemic, which saw an increase in anti-immigration rhetoric and policies.

Different social media platforms serve unique roles and attract diverse user bases. Facebook, Twitter, and YouTube are prominent platforms in driving news consumption among American adults, each offering distinct user experiences and content focuses:

• Facebook emphasizes user networks and relationships, Twitter broadcasts posts to the general public, and YouTube focuses on audiovisual content. These differences shape how users interact with content and influence the agenda-setting capacities of each platform.
• Twitter and Facebook users tend to be younger, better educated, and more politically engaged.
• Ideological compositions vary, with Twitter leaning more liberal and YouTube being popular among conservatives.
• Interaction patterns also differ; Twitter shows higher polarization, while Facebook provides a more cross-cutting space for debate.

These differences affect how news media and policy announcements are reflected in the discourse on each platform.

Research Focus and Questions

Kim et al.’s study focuses on the issue of immigration during the first year of the COVID-19 pandemic, under Trump’s presidency. It seeks to understand how different thematic emphases on immigration were shaped by news media and public policy on social media platforms.

The study poses the following hypothesis and research questions:

• Hypothesis: News coverage and policy announcements set the public agenda on social media at the second level during the first year of the COVID-19 pandemic.
• Research Question 1: How did the attribute agenda-setting effect of news coverage and policy announcements vary by issue attribute?
• Research Question 2: How did the attribute agenda-setting effect of news coverage and policy announcements vary by social media platforms (i.e., Facebook, Twitter, and YouTube)?

By exploring these questions, the study aims to provide a comprehensive understanding of how different attributes of the immigration issue were emphasized on social media, influenced by both news media and public policy.

Methods

Kim et al.’s study used three main datasets covering the period from January 1, 2020, to January 20, 2021:

1. Social Media Posts: They collected data from Twitter, Facebook, and YouTube using Synthesio, a social media monitoring tool. Synthesio aggregates publicly available data, capturing 10% of all tweets, all posts from public Facebook pages, and all YouTube comments. We used keywords related to immigration (e.g., immigrants, illegals, DACA) and expanded our list iteratively. This resulted in 9,385,469 tweets, 2,391,750 Facebook posts, and 495,317 YouTube comments.
2. News Media Headlines: Using Media Cloud, they obtained 16,263 news stories from liberal (MSNBC, CNN, Huffington Post, New York Times) and conservative (Fox News, Breitbart, Wall Street Journal, Washington Times) outlets. These outlets were selected based on their partisan scores to represent a balanced political perspective. Keywords were adjusted to fit journalistic styles.
3. Policy Documents: From the Federal Register, they collected 445 policy documents related to immigration. After relevance coding, 73 documents focused on immigration policy announcements were included. These were classified into four categories: immigrants, economy (work visas), travel (entry restrictions), and COVID-19 related measures.

To analyze social media data:

• Topic Modeling: They applied Gibbs sampling for topic modeling on social media data, determining 180 optimal topics using perplexity and loglikelihood metrics. Topics were interpreted by all authors and grouped into eight thematic categories: Immigrants, Illegals, Travel, Economy, Border, Law Enforcement, Trump’s Immigration Policy, and Court Rulings.
• Daily Post Count: Each social media post was assigned to the topic with the highest probability, and daily counts of posts per topic were calculated for each platform.

To analyze news media headlines:

• Headline Topic Modeling: We focused on headlines to capture news attention to immigration, applying Gibbs sampling to identify 50 topics. The authors interpreted the results to develop keywords for each issue attribute.
• Time Series Creation: Using the developed keywords, headlines were classified into issue attributes, and a daily count of headlines for each attribute was recorded for both liberal and conservative outlets.

Finally, to analyze policy documents:

• Relevance Coding: Three authors coded policy documents for relevance, achieving intercoder reliability. The documents were categorized into four issue attributes: immigrants, economy, travel, and COVID-19.
• Time Series Creation: Daily counts of policy documents under each attribute were tabulated.

The authors also employed time series modeling using Prais-Winsten estimation to examine the temporal relationship between predictors (news media and policy variables) and the outcome (social media thematic emphases). This method investigates how changes in predictors from one day to the next are associated with changes in the outcome.

• Correlation Matrix: They identified a high correlation (0.84) between liberal and conservative headlines on travel, leading them to exclude the liberal travel variable from analysis. Due to data sparsity, they also dropped liberal headlines on illegal immigrants.
• Stationarity and Autocorrelation: Augmented Dickey-Fuller tests confirmed that all outcome variables were stationary. Durbin Watson statistics (ranging from 1.5 to 2.5) indicated minimal residual autocorrelation, ensuring the reliability of their regression models.

This comprehensive analysis allowed Kim et al. to understand how different sources of information and policy decisions impact public discourse on social media regarding immigration.

Findings

The study found that both liberal and conservative news media significantly impacted social media discussions, particularly regarding immigration. For example, liberal news headlines on topics like border control and court rulings consistently increased related discussions on all three platforms. Similarly, conservative news coverage of travel restrictions led to more conversations about travel-related issues.

The researchers observed both positive and negative relationships between media coverage and social media responses.

• Positive relationships were evident when an increase in news headlines on a particular issue led to more discussions on the same topic across social media platforms. For instance, liberal news coverage on border issues significantly boosted related discussions on Twitter (b = 261.88, p < .001), Facebook (b = 23.68, p < .001), and YouTube (b = 5.14, p < .001).
• Conversely, negative relationships occurred when certain news coverage reduced discussions on other topics. Conservative headlines about border control, for example, decreased conversations about travel on all three platforms.

Policy announcements also influenced social media discourse, albeit less consistently.

• Policies related to immigration increased discussions about immigrants on YouTube but had mixed effects on other platforms.
• COVID-19-related policies, however, did spur conversations about border issues on Twitter and immigrant discussions on Facebook.

The study highlighted differences in how various social media platforms responded to news coverage and policy announcements.

• In general, Facebook and YouTube appeared more responsive to news and policies compared to Twitter.
• For example, while certain news topics led to significant discussions on Facebook and YouTube, their impact on Twitter was less pronounced.

In sum, this research underscores the powerful role of news media in shaping public discourse on social media.

• It reveals that both liberal and conservative media can effectively set the agenda on social media platforms, influencing what issues people discuss.
• Additionally, the findings show that social media platforms differ in how they respond to news and policy announcements, likely due to their unique user demographics and platform characteristics.

This study advances our understanding of how media and policy announcements influence social media discussions, particularly during critical times like the COVID-19 pandemic. It highlights the need for considering platform-specific dynamics when examining the flow of information and public discourse. As social media continues to play a pivotal role in public discussions, recognizing these influences can help us better navigate and understand the digital information landscape.

What Can You Do?

Understanding the nuances of agenda-setting theory and its application on social media is crucial for both media professionals and consumers. Here are practical steps to address the main issues discussed in the study:

1. Media Literacy Programs: Implement educational programs to enhance media literacy, helping individuals critically analyze news sources and identify biases.
2. Platform Accountability: Advocate for social media platforms to increase transparency about their algorithms and how they prioritize certain topics.
3. Diverse News Consumption: Encourage consuming a broad spectrum of news sources to gain a more balanced understanding of issues.
4. Fact-Checking Initiatives: Support fact-checking organizations to verify the information circulating on social media.
5. Policy Engagement: Engage with policymakers to ensure that public policies are informed by a broad array of perspectives, not just dominant media narratives.
6. Research Support: Fund and support research into media effects on public perception to better understand and mitigate undue influence.
7. Public Awareness Campaigns: Run campaigns to raise public awareness about the effects of media on public opinion.
8. Ethical Journalism: Promote ethical journalism standards that emphasize unbiased and comprehensive coverage of issues.
9. Content Moderation: Encourage social media platforms to develop robust content moderation policies to curb misinformation.
10. Community Building: Foster online communities that encourage respectful and informed discussions about current issues.

Have you seen the media driving topics of discussion on your own social media feeds? What do you think can be done to encourage more diverse, accurate perspectives on social media? Share your ideas in the comments!

Stay informed and inspired! Subscribe to our newsletter, “Society This Week,” and get a free copy of our exclusive report, “Peaceful, Powerful: 225 Nonviolent Tactics for Impactful Activism!”