Every instrument has an intrinsic accuracy, and the results of measurements must acknowledge those errors.

Error propagation, or uncertainty analysis, is one of those skills that every experimentalist needs to master.  Biologist, chemist, and physicist—when they report a result (an “answer”), they must also report how accurate that result is, and how confident they are in the measurement.  Essentially, if they get an answer of “5″, does that mean the true value is somewhere between 4.99 and 5.01, or between 2 and 8?  As with every skill, there are levels of mastery, and everyone needs to start somewhere.  The folks at Wellesley college have provided a great introduction to the topic, using analytical balance measurements as an example.