Analyzing the Impact of Social Media Buy Recommendations on Saudi Stock Market Reactions: A Bert-Based Sentiment and Network Analysis Approach

<p>Social media has become part of the daily rhythm of financial markets, its actual influence on prices remains debated. This dissertation examines whether buy recommendations posted by Saudi-focused social media analysts on Twitter/X are associated with short-term stock market reactions in the Tadawul exchange. The central question is straightforward: when influential online analysts issue concentrated buy signals, do prices and trading activity respond in measurable ways? To address this, I combine multiple empirical strategies in a triangulated research design. Fixed-effects panel regressions estimate the association between sentiment measures, analyst exposure, abnormal returns and volume while accounting for firm-level and time-specific factors. Propensity score matching is used to compare exposed and non-exposed stock-days with similar observable characteristics, helping reduce selection concerns and strengthening causal interpretation. An event-study framework aligns returns around recommendation bursts to examine dynamic price adjustments within clearly defined exposure windows, helping isolate the timing of market responses. Each method approaches the question from a slightly different angle. The patterns suggest that analyst activity is strongly linked to short-lived increases in trading volume and modest return movements. Effects appear concentrated around exposure shocks and tend to dissipate within days, which consistent with attention-driven responses. Pre-event drift hints that analysts often post during periods of existing market stress, complicating clean causal interpretation. To complement the quantitative analysis, qualitative interviews with retail investors explore how credibility, repetition, and network visibility are perceived to influence decision-making. Investors describe using informal mental checklists before acting on recommendations. Their accounts help contextualize the statistical findings and suggest that social media effects operate through attention and trust. Overall, the findings suggest that social media recommendations can redirect attention and temporarily intensify trading, especially in retail-driven markets. The influence appears situational, shaped by credibility, timing, and existing sentiment. These patterns are consistent with a short-term causal influence that is conditioned by market context and investor attention.</p>

Name thesis:degree_name

Information Systems and Technology, PhD

Level thesis:degree_level

Restricted to Claremont Colleges Dissertation

Discipline thesis:degree_discipline

Center for Information Systems and Technology

Year dc:date.available

2026

Author dc:creator

• Alharbi, Fayez

Contributors dc:contributor

• Wallace Chipidza
• Chinazunwa Uwaoma

Repository record dc:identifier

https://scholarship.claremont.edu/cgu_etd/1084

OAI identifier oai:identifier

oai:scholarship.claremont.edu:cgu_etd-2106