In the context where your measurements allow you to measure discrepancies a seventh of that error, that’s a huge error (for the purpose of THIS experiment). The

In the context where your measurements allow you to measure discrepancies a seventh of that error, that’s a huge error (for the purpose of THIS experiment). There might be some experiments where the difference doesn’t matter at all, or other experiments where the dependence is linear, so the answer also shifts by 0.1%. But the dependence could just as easily be some complicated nonlinear function which leads to a large discrepancy compared to the measurement precision.