The correct answer is $\boxed{\text{B. 25}}$.
The residual error is the difference between the observed value and the predicted value. In this case, the observed value is unknown, but the predicted value is 50 and the residual error is 25. Therefore, the observed value must be 50 – 25 = $\boxed{\text{25}}$.
Option A is incorrect because 100 is greater than the predicted value of 50.
Option C is incorrect because 50 is the predicted value, not the observed value.
Option D is incorrect because 75 is greater than the observed value of 25.