Behavioural biases are thought to be important influencers in explainingspeculative investment, leading to irrational stock market decision-making. This study aimsto assess the influence of behavioral biases called herding and social media scams on stockmarket investment decisions in Bangladesh with a special focus on the mediating role ofoverconfidence. Based on the Behavioural Bias theory and grounded in a positivist researchparadigm, the study adopts a quantitative methodology. A convenient sampling technique isemployed to collect data through a structured questionnaire administered to 700 investors inthe stock markets of Bangladesh, achieving a response rate of 60.29% with 395 respondents.The collected data were analyzed using Partial Least Squares Structural Equation Modeling(PLS-SEM). The study reveals that investment decisions are greatly influenced by socialmedia scams and herding biases. Similarly, social media and herding bias greatly contributeto the development of overconfidence. Furthermore, irrational investing decisions arepartially mediated by overconfidence. The study provides practical implications forpolicymakers and financial institutions to design behavioral interventions and educationalprograms aimed at mitigating the adverse effects of such behavioral biases. Byunderstanding common cognitive biases, advisors can help clients avoid following the crowdor making decisions based on social influence. Advisors can also play a key role inpromoting more independent and evidence-based investment strategies.