The thermal maturity of shale is often measured by vitrinite reflectance (VRo). VRo measurements for the Devonian–Mississippian black shale source rocks evaluated herein predicted thermal immaturity in areas where associated reservoir rocks are oil-producing. This limitation of the VRo method led to the current evaluation of Raman spectroscopy as a suitable alternative for developing correlations between thermal maturity and Raman spectra. In this study, Raman spectra of Devonian–Mississippian black shale source rocks were regressed against measured VRo or sample-depth. Attempts were made to develop quantitative correlations of thermal maturity. Using sample-depth as a proxy for thermal maturity is not without limitations as thermal maturity as a function of depth depends on thermal gradient, which can vary through time, subsidence rate, uplift, lack of uplift, and faulting. Correlations between Raman data and vitrinite reflectance or sample-depth were quantified by peak-fitting the spectra. Various peak-fitting procedures were evaluated to determine the effects of the number of peaks and maximum peak widths on correlations between spectral metrics and thermal maturity. Correlations between D-frequency, G-band full width at half maximum (FWHM), and band separation between the G- and D-peaks and thermal maturity provided some degree of linearity throughout most peak-fitting assessments; however, these correlations and those calculated from the G-frequency, D/G FWHM ratio, and D/G peak area ratio also revealed a strong dependence on peak-fitting processes. This dependency on spectral analysis techniques raises questions about the validity of peak-fitting, particularly given the amount of subjective analyst involvement necessary to reconstruct spectra. This research shows how user interpretation and extrapolation affected the comparability of different samples, the accuracy of generated trends, and therefore, the potential of the Raman spectral method to become an industry benchmark as a thermal maturity probe. A Raman method devoid of extensive operator interaction and data manipulation is quintessential for creating a standard method.