Impact of Social Media Algorithms on Local News Trust and Engagement

Thesis Abstract

The rapid proliferation of social media platforms has transformed how local news is produced, distributed, and perceived, raising concerns about the reliability of information and the engagement it garners among community audiences. This study addresses the problem that algorithmic curation on social media may distort trust in local news and alter audience engagement patterns, potentially undermining civic participation at the local level. The aim is to examine how algorithmic personalization influences verified local news trust, perceived credibility, and engagement behaviors among residents in urban and peri-urban communities. Specific objectives are (1) to assess the relationship between exposure to local news through algorithm-driven feeds and trust in local news sources; (2) to evaluate how algorithmic ranking affects perceived credibility and willingness to share local news content; (3) to identify differential effects of personalization on engagement metrics across demographic groups (age, education, and income); (4) to explore moderating roles of media literacy and prior local news consumption; and (5) to propose practical guidelines for local news outlets and platform designers to enhance trust and constructive engagement. The study adopts a mixed-methods design integrating a cross-sectional survey and in-depth interviews. The population comprises adults aged 18 and above residing in three metropolitan areas with diverse media ecosystems. A stratified random sample of 1,200 respondents will be surveyed, with quotas to ensure representation by age, gender, education, and income. Instrumentation includes a structured questionnaire measuring perceived local news credibility (validated scales adapted from existing trust in news scales), intent to share local news, frequency of use of social media for local news, and media literacy. In addition, a subset of 40 survey participants will be purposively selected for semi-structured interviews to gain nuanced insights into experiential dimensions of algorithmic curation and trust dynamics. Data collection will be conducted over six months, with online survey administration and in-depth interviews conducted via videoconference. Quantitative data will be analyzed using structural equation modeling (SEM) to test a theoretical model linking algorithmic exposure, perceived credibility, trust in local news, and engagement outcomes, while controlling for demographic covariates. Multigroup SEM will assess differential effects across demographic strata. Regression-based mediation analyses will examine whether perceived credibility mediates the relationship between algorithmic exposure and sharing/engagement intentions. Reliability and validity will be established through Cronbach’s alpha, confirmatory factor analysis, and measurement invariance testing. Qualitative data from interviews will be analyzed using thematic analysis to identify themes related to algorithmic transparency, perceived bias, confirmation of local identity, and civic engagement implications. Triangulation will integrate findings from both strands to corroborate the proposed relationships and interpret divergent patterns. The expected findings include (a) higher exposure to algorithm-curated local news is associated with lower perceived credibility and trust, particularly among participants with limited media literacy; (b) credibility acts as a mediator between algorithmic exposure and engagement behaviors such as commenting, liking, and sharing, with stronger effects for participants who value local identity; (c) demographic factors, notably age and education, modulate the strength of relationships between exposure, trust, and engagement; and (d) participants report a demand for greater transparency about how local news is prioritized and a need for diverse local perspectives to counterbalance homogenized content. The study contributes to knowledge by integrating theories of agenda setting, cognitive misperception, and the dual-process model of persuasion to illuminate how algorithmic personalization intersects with trust formation in local news ecosystems. It also provides empirical evidence on the social consequences of algorithmic curation for local democracy and public sphere participation. Based on the findings, practical recommendations will be advanced for local news organizations to foster trust and responsible engagement, including transparent disclosure of curation criteria, development of editorial interventions to counteract echo chambers, and media literacy initiatives tailored to local communities. Platform designers will be advised to incorporate user controls over personalization settings, and to promote diversified local news streams to enhance credibility and participatory engagement. The study concludes that while algorithmic personalization can amplify engagement, it may simultaneously erode trust in local news unless accompanied by transparent practices and targeted media literacy interventions.