This article describes how a multiple-choice single-response (MCSR) question was built, validated and calibrated to measure learning of the Bernoulli's law in a population of engineering students. We have used the test's classical theory to reliably return the instrument using three indexes which adjust only the questions and two that adjust the entire test. Similarly, outlines some alternative conceptions found in populations intervened to build both the questions and distracters. We report the test because it can serve as a research tool in other populations, but it only evaluates the conceptual dimension of Bernoulli's law for fluids. The methodology reported can be very suitable for creating and validating assessment tools that measures other Physics concepts in Latin-American.