Thermal Response Tests (TRTs) are widely used to determine the effective ground thermal conductivity (λeff) and borehole thermal resistance (Rb) for the design of ground-source heat pump systems. However, despite established recommendations regarding TRT duration, the influence of the length and position of the data analysis window, particularly under different levels of operational stability, remains insufficiently quantified using field measurements. This study addresses this gap by analyzing three field-based TRTs conducted in boreholes located within the same geological site using the Infinite Line Source (ILS) model and eight different time intervals. Across all analyzed intervals, λeff ranged from 2.00 to 3.47 W/(m·K) for TRT1, 1.88 to 2.76 W/(m·K) for TRT2, and 2.14 to 2.62 W/(m·K) for TRT3. The corresponding predicted average ground thermal yields ranged from 38.0 to 62.3, 36.0 to 50.5, and 40.3 to 48.2 W/m, respectively. When only analysis intervals starting at 10 h or later were considered, the λeff ranges were reduced to 2.80–3.47, 1.88–2.76, and 2.44–2.62 W/(m·K), respectively. The ILS model reproduced the measured fluid temperatures with MAPE values ranging from 0.42% to 2.43%. The results demonstrate that excluding early-time data or satisfying the minimum Fourier number criterion alone does not ensure consistent parameter estimation. In particular, localized operational disturbances can substantially affect the estimated thermal parameters even at later stages of a test. The study provides field-based quantitative evidence that both analysis-window selection and test stability are critical for reliable ILS-based TRT interpretation and for the subsequent use of the estimated parameters in borehole heat exchanger design.