It has long been known that chronic exposures to high concentrations of respirable coal mine dust can lead to the development of lung diseases such as Coal Worker's Pneumoconiosis, commonly referred to as "black lung", and silicosis. Since the mid-1990s, an alarming resurgence of diseases has been documented in central Appalachia, where underground mining often necessitates significant extraction of rock strata along with the thin seams of coal. These circumstances have prompted concern over if or how changing dust composition might be a factor in contemporary disease prevalence. Until now, the total mass concentration and quartz mass fraction of respirable dust have been regulated and monitored in US coal mines. Unfortunately, however, these two metrics alone do not paint a full picture of dust composition. Earlier work in the author's research group established a preliminary thermogravimetric analysis (TGA) method for coal mine dust. The method is intended to allow estimation of three key mass fractions of the dust from separate sources: coal from the coal strata being mined; non-carbonate minerals from the rock strata being mined or drilled; and carbonates that are primarly sourced from application of rock dust products to the mine floor or ribs. However, accuracy of the preliminary method was substantially limited by poor dust recovery from the fibrous filter media used for sample collection. This thesis includes two studies: The first study aims to establish an improved TGA method. It uses smooth polycarbonate (PC) filters for dust sampling and a modified thermal ramping routine. The method is verified using laboratory-generated respirable dust samples. In the second study, the improved TGA method is used to analyze 75 respirable mine dust samples, collected in 15 US mines. Replicate samples are also analyzed by scanning electron microscopy using energy dispersive X-ray (SEM-EDX). TGA and SEM-EDX results are compared to gain insights regarding the analytical methods and general trends in dust composition within and between mines. / Master of Science / It has long been known that chronic exposures to excessive respirable coal mine dust can lead to the development of lung diseases such as Coal Worker’s Pneumoconiosis (“Black Lung”) and silicosis. Disease rates in central Appalachia have shown an alarming and unexpected increase since the mid-1990s, despite declining dust concentrations evident from regulatory compliance monitoring data. Clearly, there is a need to better understand coal mine dust composition, which will require additional analytical methods. Thermogravimetric analysis (TGA) has been proposed as one possible method, because it should allow estimation of three key dust components from separate sources: coal from the coal strata being mined; non-carbonate minerals from the rock strata being mined or drilled; and carbonates from application of rock dust products to the mine floor and ribs. However, preliminary work with TGA showed limited accuracy, mostly due to sampling materials. In this thesis, two studies were performed. The first study aims to establish an improved TGA method using smooth, polycarbonate (PC) filters. The second study demonstrates the method on a large number of mine dust samples, and compares the results to those gained by an alternative method that uses electron microscopy.