The Galileo High Accuracy Service (HAS) provides State Space Representation (SSR) corrections for satellite orbits, clocks and biases, enabling high-accuracy GNSS applications. While several studies have assessed HAS-based positioning, fewer works have focused on a rigorous statistical evaluation of HAS-corrected orbits themselves. This paper presents a comparative analysis of Galileo broadcast (BRDC) and HAS-corrected orbits using IGS SP3 products as reference. For 23 Galileo satellites over a 20-day period, 3D orbit errors at the antenna phase center are computed and analysed through a block-wise paired framework combining classical tests and block bootstrap confidence intervals. Results show a systematic and statistically significant RMS reduction for all satellites, with relative improvements typically between 4–6% and up to about 8.4%. The proposed methodology provides a robust template for evaluating GNSS orbit correction services under temporal correlation.