In representative democracy, it is crucial to include the perspectives of those governed in policy making. To analyze representation, research often links public policy preferences with legislators' stances through surveys and votes. However, the scholarship lacks effective methods to assess whether substantive policy ideas of the public gain lawmakers' attention. This study combines Reddit discussions on policy issues with U.S. House of Representatives' hearing transcripts from 2005 to 2022 to develop an innovative LLM-driven argument detection and stance classification framework called WIBA (``What is Being Argued''). By applying WIBA, we construct the new measurements of representation by identifying similar arguments made in both Reddit and hearings and show how the pattern of representation varies across policy issues (partisan vs. non-partisan), legislators, and Reddit communities. Our proposed approach makes both substantive and methodological contributions by introducing a new approach to conceptualize and measure representation at the argument level that is more nuanced and granular than any of the existing methods to measure this concept, providing a deeper understanding of representation.