A test consists of 910 true or false questions. If the student guesses on each question, what is the expected standard deviation of the number of correct answers?
Such a variable would be well-modeled by a Normal distribution with mean
These are approximations and we are not making a distinction between these ranges including the endpoints or not. The "half-unit (continuity) correction " could be used along with a table of Normal probabilities to try to make these estimates more precise.