The original TRAPPIST-1 e spectrum and the corrected TRAPPIST-1 e spectrum for observation 15, where “corrected” means dividing by the TRAPPIST-1 b spectrum and then multiplying by the median transit depth of the TRAPPIST-1 b spectrum to return them to the same scale. Plotted on top are GP and flat line fits to the data. The number in the top right corner is the Δln Z between the flat line and GP models, such that a negative number prefers a GP and a positive number prefers a flat line. On the left are the results from the independent light-curve fitting described in Section 3.1, and on the right are the results from the consistent light-curve fitting test carried out using the setup from the N.A. reduction. For all reductions and both cases, the corrected spectrum more strongly prefers a flat line over the GP model than the original spectrum, which suggests that the corrected spectra are more consistent with a flat line and are corrected, at least in part, for the stellar contamination signal. The shape of the GP model for the corrected spectra differs somewhat for each reduction in the independent light-curve fitting case, but the shape looks very similar across the reductions for the consistent light-curve fitting case. For all cases, a flat line is always more strongly preferred after the spectroscopic ratio is taken.